refactor(server): simplify modal process lifecycle

feat(server): add modal environment driver
feat(core): export environment conformance suite
2026-08-07 17:49:53 -04:00 · 2026-08-07 13:21:44 -04:00 · 2026-08-07 13:16:08 -04:00 · 2026-08-07 13:16:05 -04:00
31 changed files with 498 additions and 994 deletions
@@ -679,6 +679,7 @@
        "@opencode-ai/util": "workspace:*",
        "drizzle-orm": "catalog:",
        "effect": "catalog:",
+        "modal": "0.9.0",
      },
      "devDependencies": {
        "@tsconfig/bun": "catalog:",
@@ -1512,6 +1513,18 @@

    "@capsizecss/unpack": ["@capsizecss/unpack@2.4.0", "", { "dependencies": { "blob-to-buffer": "^1.2.8", "cross-fetch": "^3.0.4", "fontkit": "^2.0.2" } }, "sha512-GrSU71meACqcmIUxPYOJvGKF0yryjN/L1aCuE9DViCTJI7bfkjgYDPD1zbNDcINJwSSP6UaBZY9GAbYDO7re0Q=="],

+    "@cbor-extract/cbor-extract-darwin-arm64": ["@cbor-extract/cbor-extract-darwin-arm64@2.2.2", "", { "os": "darwin", "cpu": "arm64" }, "sha512-ZKZ/F8US7JR92J4DMct6cLW/Y66o2K576+zjlEN/MevH70bFIsB10wkZEQPLzl2oNh2SMGy55xpJ9JoBRl5DOA=="],
+
+    "@cbor-extract/cbor-extract-darwin-x64": ["@cbor-extract/cbor-extract-darwin-x64@2.2.2", "", { "os": "darwin", "cpu": "x64" }, "sha512-32b1mgc+P61Js+KW9VZv/c+xRw5EfmOcPx990JbCBSkYJFY0l25VinvyyWfl+3KjibQmAcYwmyzKF9J4DyKP/Q=="],
+
+    "@cbor-extract/cbor-extract-linux-arm": ["@cbor-extract/cbor-extract-linux-arm@2.2.2", "", { "os": "linux", "cpu": "arm" }, "sha512-tNg0za41TpQfkhWjptD+0gSD2fggMiDCSacuIeELyb2xZhr7PrhPe5h66Jc67B/5dmpIhI2QOUtv4SBsricyYQ=="],
+
+    "@cbor-extract/cbor-extract-linux-arm64": ["@cbor-extract/cbor-extract-linux-arm64@2.2.2", "", { "os": "linux", "cpu": "arm64" }, "sha512-wfqgzqCAy/Vn8i6WVIh7qZd0DdBFaWBjPdB6ma+Wihcjv0gHqD/mw3ouVv7kbbUNrab6dKEx/w3xQZEdeXIlzg=="],
+
+    "@cbor-extract/cbor-extract-linux-x64": ["@cbor-extract/cbor-extract-linux-x64@2.2.2", "", { "os": "linux", "cpu": "x64" }, "sha512-rpiLnVEsqtPJ+mXTdx1rfz4RtUGYIUg2rUAZgd1KjiC1SehYUSkJN7Yh+aVfSjvCGtVP0/bfkQkXpPXKbmSUaA=="],
+
+    "@cbor-extract/cbor-extract-win32-x64": ["@cbor-extract/cbor-extract-win32-x64@2.2.2", "", { "os": "win32", "cpu": "x64" }, "sha512-dI+9P7cfWxkTQ+oE+7Aa6onEn92PHgfWXZivjNheCRmTBDBf2fx6RyTi0cmgpYLnD1KLZK9ZYrMxaPZ4oiXhGA=="],
+
    "@chevrotain/types": ["@chevrotain/types@11.1.2", "", {}, "sha512-U+HFai5+zmJCkK86QsaJtoITlboZHBqrVketcO2ROv865xfCMSFpELQoz1GkX5GzME8pTa+3kbKrZHQtI0gdbw=="],

    "@clack/core": ["@clack/core@1.0.0-alpha.1", "", { "dependencies": { "picocolors": "^1.0.0", "sisteransi": "^1.0.5" } }, "sha512-rFbCU83JnN7l3W1nfgCqqme4ZZvTTgsiKQ6FM0l+r0P+o2eJpExcocBUWUIwnDzL76Aca9VhUdWmB2MbUv+Qyg=="],
@@ -1720,6 +1733,10 @@

    "@graphql-typed-document-node/core": ["@graphql-typed-document-node/core@3.2.0", "", { "peerDependencies": { "graphql": "^0.8.0 || ^0.9.0 || ^0.10.0 || ^0.11.0 || ^0.12.0 || ^0.13.0 || ^14.0.0 || ^15.0.0 || ^16.0.0 || ^17.0.0" } }, "sha512-mB9oAsNCm9aM3/SOv4YtBMqZbYj10R7dkq8byBqxGY/ncFwhf2oQzMV+LCRlWoDSEBJ3COiR1yeDvMtsoOsuFQ=="],

+    "@grpc/grpc-js": ["@grpc/grpc-js@1.14.4", "", { "dependencies": { "@grpc/proto-loader": "^0.8.0", "@js-sdsl/ordered-map": "^4.4.2" } }, "sha512-k9Dj3DV/itK9D06Y8f190Qgop7/Ui+D0njFV3LHMPwPT75DpXLQohE9Wmz0QElrJnzsjB7KPWiKJbOl7IPDArQ=="],
+
+    "@grpc/proto-loader": ["@grpc/proto-loader@0.8.1", "", { "dependencies": { "lodash.camelcase": "^4.3.0", "long": "^5.0.0", "protobufjs": "^7.5.5", "yargs": "^17.7.2" }, "bin": { "proto-loader-gen-types": "build/bin/proto-loader-gen-types.js" } }, "sha512-wtF6h+DY6M3YaDBPAmvuuA6jV8Sif9MjtOI5euKFWRgCDl5PeDpPsHR9u2l6St5ceY8AZgoNDww5+HvEsXFsGg=="],
+
    "@happy-dom/global-registrator": ["@happy-dom/global-registrator@20.0.11", "", { "dependencies": { "@types/node": "^20.0.0", "happy-dom": "^20.0.11" } }, "sha512-GqNqiShBT/lzkHTMC/slKBrvN0DsD4Di8ssBk4aDaVgEn+2WMzE6DXxq701ndSXj7/0cJ8mNT71pM7Bnrr6JRw=="],

    "@hono/node-server": ["@hono/node-server@1.19.15", "", { "peerDependencies": { "hono": "^4" } }, "sha512-Za2ai6TLdKjUvnur+eenO6nuYYipVAEhyCAdaV8IRvmU9kK8crOZUSYvIXn72E4f8fJqyAbpcJuTsYYmZp9Deg=="],
@@ -1814,6 +1831,8 @@

    "@jridgewell/trace-mapping": ["@jridgewell/trace-mapping@0.3.31", "", { "dependencies": { "@jridgewell/resolve-uri": "^3.1.0", "@jridgewell/sourcemap-codec": "^1.4.14" } }, "sha512-zzNR+SdQSDJzc8joaeP8QQoCQr8NuYx2dIIytl1QeBEZHJ9uW6hebsrYgbz8hJwUQao3TWCMtmfV8Nu1twOLAw=="],

+    "@js-sdsl/ordered-map": ["@js-sdsl/ordered-map@4.4.2", "", {}, "sha512-iUKgm52T8HOE/makSxjqoWhe95ZJA1/G1sYsGev2JDKUSS14KAgg1LHb+Ba+IPow0xflbnSkOsZcO08C7w1gYw=="],
+
    "@js-temporal/polyfill": ["@js-temporal/polyfill@0.5.1", "", { "dependencies": { "jsbi": "^4.3.0" } }, "sha512-hloP58zRVCRSpgDxmqCWJNlizAlUgJFqG2ypq79DCvyv9tHjRYMDOcPFjzfl/A1/YxDvRCZz8wvZvmapQnKwFQ=="],

    "@jsx-email/all": ["@jsx-email/all@2.2.3", "", { "dependencies": { "@jsx-email/body": "1.0.2", "@jsx-email/button": "1.0.4", "@jsx-email/column": "1.0.3", "@jsx-email/container": "1.0.2", "@jsx-email/font": "1.0.3", "@jsx-email/head": "1.0.2", "@jsx-email/heading": "1.0.2", "@jsx-email/hr": "1.0.2", "@jsx-email/html": "1.0.2", "@jsx-email/img": "1.0.2", "@jsx-email/link": "1.0.2", "@jsx-email/markdown": "2.0.4", "@jsx-email/preview": "1.0.2", "@jsx-email/render": "1.1.1", "@jsx-email/row": "1.0.2", "@jsx-email/section": "1.0.2", "@jsx-email/tailwind": "2.4.4", "@jsx-email/text": "1.0.2" }, "peerDependencies": { "react": "^18.2.0" } }, "sha512-OBvLe/hVSQc0LlMSTJnkjFoqs3bmxcC4zpy/5pT5agPCSKMvAKQjzmsc2xJ2wO73jSpRV1K/g38GmvdCfrhSoQ=="],
@@ -3298,6 +3317,8 @@

    "abort-controller": ["abort-controller@3.0.0", "", { "dependencies": { "event-target-shim": "^5.0.0" } }, "sha512-h8lQ8tacZYnR3vNQTgibj+tODHI5/+l06Au2Pcriv/Gmet0eaj4TwWH41sO9wnHDiQsEj19q0drzdWdeAHtweg=="],

+    "abort-controller-x": ["abort-controller-x@0.5.0", "", {}, "sha512-yTt9CI0x+nRfX6BFMenEGP8ooPvErGH6AbFz20C2IeOLIlDsrw/VHpgne3GsCEuTA410IiFiaLVFKmgM4bKEPQ=="],
+
    "accepts": ["accepts@2.0.0", "", { "dependencies": { "mime-types": "^3.0.0", "negotiator": "^1.0.0" } }, "sha512-5cvg6CtKwfgdmVqY1WIiXKc3Q1bkRqGLi+2W/6ao+6Y7gu/RCwRuAhGEzh5B4KlszSuTLgZYuqFqo5bImjNKng=="],

    "acorn": ["acorn@8.15.0", "", { "bin": { "acorn": "bin/acorn" } }, "sha512-NZyJarBfL7nWwIq+FDL6Zp/yHEhePMNnnJ0y3qfieCrmNvYct8uvtiV41UvlSe6apAfk0fY1FbWx+NwfmpvtTg=="],
@@ -3538,6 +3559,10 @@

    "caniuse-lite": ["caniuse-lite@1.0.30001806", "", {}, "sha512-72Cuvd95zbSYPKq6Fhg8eDJRlzgWDf7/mtoZv6Qe/DYNCEBdNxoA3+rZAU2ZhGCpZlns3EssFavaZomckT5Uuw=="],

+    "cbor-extract": ["cbor-extract@2.2.2", "", { "dependencies": { "node-gyp-build-optional-packages": "5.1.1" }, "optionalDependencies": { "@cbor-extract/cbor-extract-darwin-arm64": "2.2.2", "@cbor-extract/cbor-extract-darwin-x64": "2.2.2", "@cbor-extract/cbor-extract-linux-arm": "2.2.2", "@cbor-extract/cbor-extract-linux-arm64": "2.2.2", "@cbor-extract/cbor-extract-linux-x64": "2.2.2", "@cbor-extract/cbor-extract-win32-x64": "2.2.2" }, "bin": { "download-cbor-prebuilds": "bin/download-prebuilds.js" } }, "sha512-hlSxxI9XO2yQfe9g6msd3g4xCfDqK5T5P0fRMLuaLHhxn4ViPrm+a+MUfhrvH2W962RGxcBwEGzLQyjbDG1gng=="],
+
+    "cbor-x": ["cbor-x@1.6.5", "", { "optionalDependencies": { "cbor-extract": "^2.2.2" } }, "sha512-yO64CxnSh6kp+pHNRK9IfwnMvCB+c8HvmUjQY/9l9YRF0/cAPka/tUHLwS64QqUpFCq3/OtbKziVJYXH2EaRig=="],
+
    "ccount": ["ccount@2.0.1", "", {}, "sha512-eyrF0jiFpY+3drT6383f1qhkbGsLSifNAjA61IUjZjmLCWjItY6LB9ft9YhoDgwfmclB2zhu51Lc7+95b8NRAg=="],

    "chai": ["chai@5.3.3", "", { "dependencies": { "assertion-error": "^2.0.1", "check-error": "^2.1.1", "deep-eql": "^5.0.1", "loupe": "^3.1.0", "pathval": "^2.0.0" } }, "sha512-4zNhdJD/iOjSH0A05ea+Ke6MU5mmpQcbQsSOkgdaUMJ9zTlDTD/GYlwohmIE2u0gaxHYiVHEn1Fw9mZ/ktJWgw=="],
@@ -4570,6 +4595,8 @@

    "lodash-es": ["lodash-es@4.18.1", "", {}, "sha512-J8xewKD/Gk22OZbhpOVSwcs60zhd95ESDwezOFuA3/099925PdHJ7OFHNTGtajL3AlZkykD32HykiMo+BIBI8A=="],

+    "lodash.camelcase": ["lodash.camelcase@4.3.0", "", {}, "sha512-TwuEnCnxbc3rAvhf/LbG7tJUDzhqXyFnv3dtzLOPgCG/hODL7WFnsbwktkD7yUV0RrreP/l1PALq/YSg6VvjlA=="],
+
    "lodash.escaperegexp": ["lodash.escaperegexp@4.1.2", "", {}, "sha512-TM9YBvyC84ZxE3rgfefxUWiQKLilstD6k7PTGt6wfbtXF8ixIJLOL3VYyV/z+ZiPLsVxAsKAFVwWlWeb2Y8Yyw=="],

    "lodash.includes": ["lodash.includes@4.3.0", "", {}, "sha512-W3Bx6mdkRTGtlJISOvVD/lbqjTlPPUDTMnlXZFnVwi9NKJ6tiAk6LVdlhZMm17VZisqhKcgzpO5Wz91PCt5b0w=="],
@@ -4796,6 +4823,8 @@

    "mkdirp": ["mkdirp@0.5.6", "", { "dependencies": { "minimist": "^1.2.6" }, "bin": { "mkdirp": "bin/cmd.js" } }, "sha512-FP+p8RB8OWpF3YZBCrP5gtADmtXApB5AMLn+vdyA+PyxCjrCs00mjyUozssO33cwDeT3wNGdLxJ5M//YqtHAJw=="],

+    "modal": ["modal@0.9.0", "", { "dependencies": { "cbor-x": "^1.6.0", "long": "^5.3.1", "nice-grpc": "^2.1.12", "protobufjs": "^7.5.0", "smol-toml": "^1.3.3", "uuid": "^11.1.0" } }, "sha512-kCXcdJkhbJorf/q/6T9Wdlg6in9JmRnCNQnV6rVBMyeqNV/iXI6BYk4IzY4cvZ6dbauNeDMjk/Q08cbxvoIaXg=="],
+
    "morphdom": ["morphdom@2.7.8", "", {}, "sha512-D/fR4xgGUyVRbdMGU6Nejea1RFzYxYtyurG4Fbv2Fi/daKlWKuXGLOdXtl+3eIwL110cI2hz1ZojGICjjFLgTg=="],

    "motion": ["motion@12.34.5", "", { "dependencies": { "framer-motion": "^12.34.5", "tslib": "^2.4.0" }, "peerDependencies": { "@emotion/is-prop-valid": "*", "react": "^18.0.0 || ^19.0.0", "react-dom": "^18.0.0 || ^19.0.0" }, "optionalPeers": ["@emotion/is-prop-valid", "react", "react-dom"] }, "sha512-N06NLJ9IeBHeielRqIvYvjPfXuRdyTxa+9++BgpGa+hY2D7TcMkI6QzV3jaRuv0aZRXgMa7cPy9YcBUBisPzAQ=="],
@@ -4834,6 +4863,10 @@

    "nf3": ["nf3@0.1.12", "", {}, "sha512-qbMXT7RTGh74MYWPeqTIED8nDW70NXOULVHpdWcdZ7IVHVnAsMV9fNugSNnvooipDc1FMOzpis7T9nXJEbJhvQ=="],

+    "nice-grpc": ["nice-grpc@2.1.16", "", { "dependencies": { "@grpc/grpc-js": "^1.14.0", "abort-controller-x": "^0.5.0", "nice-grpc-common": "^2.0.3" } }, "sha512-Cl3Pn00212Hl8/U6bpgMxmhZj5lyv3nWoJov4cd3FjWarktrMHP4DNvSjCnDwkMWYx4W1tyscEia4JX6Y4GVCQ=="],
+
+    "nice-grpc-common": ["nice-grpc-common@2.0.3", "", { "dependencies": { "ts-error": "^1.0.6" } }, "sha512-MEhnD3JMah0mgyivpb9hpRDbOBuXBxI/TVO+OK1h6rC97WM42HsPMR+zzRNQ0C5BqYJTw1nyWiQRD0DucO+pjQ=="],
+
    "nitro": ["nitro@3.0.1-alpha.1", "", { "dependencies": { "consola": "^3.4.2", "crossws": "^0.4.1", "db0": "^0.3.4", "h3": "2.0.1-rc.5", "jiti": "^2.6.1", "nf3": "^0.1.10", "ofetch": "^2.0.0-alpha.3", "ohash": "^2.0.11", "oxc-minify": "^0.96.0", "oxc-transform": "^0.96.0", "srvx": "^0.9.5", "undici": "^7.16.0", "unenv": "^2.0.0-rc.24", "unstorage": "^2.0.0-alpha.4" }, "peerDependencies": { "rolldown": "*", "rollup": "^4", "vite": "^7", "xml2js": "^0.6.2" }, "optionalPeers": ["rolldown", "rollup", "vite", "xml2js"], "bin": { "nitro": "dist/cli/index.mjs" } }, "sha512-U4AxIsXxdkxzkFrK0XAw0e5Qbojk8jQ50MjjRBtBakC4HurTtQoiZvF+lSe382jhuQZCfAyywGWOFa9QzXLFaw=="],

    "nlcst-to-string": ["nlcst-to-string@4.0.0", "", { "dependencies": { "@types/nlcst": "^2.0.0" } }, "sha512-YKLBCcUYKAg0FNlOBT6aI91qFmSiFKiluk655WzPF+DDMA02qIyy8uiRqI8QXtcFpEvll12LpL5MXqEmAZ+dcA=="],
@@ -5632,6 +5665,8 @@

    "ts-dedent": ["ts-dedent@2.3.0", "", {}, "sha512-JfJeIHke7y2egdGGgRAvpCwYFUsHlM2gPcrVOxFkznt/4uzQ7HFmvE63iFHVLBJNDuyDOQgijDK/tXH/f6Msjg=="],

+    "ts-error": ["ts-error@1.0.6", "", {}, "sha512-tLJxacIQUM82IR7JO1UUkKlYuUTmoY9HBJAmNWFzheSlDS5SPMcNIepejHJa4BpPQLAcbRhRf3GDJzyj6rbKvA=="],
+
    "ts-interface-checker": ["ts-interface-checker@0.1.13", "", {}, "sha512-Y/arvbn+rrz3JCKl9C4kVNfTfSm2/mEp5FSz5EsZSANGPSlQrpRI5M4PKF+mJnE52jOO90PnPSc3Ur3bTQw0gA=="],

    "tsconfck": ["tsconfck@3.1.6", "", { "peerDependencies": { "typescript": "^5.0.0" }, "optionalPeers": ["typescript"], "bin": { "tsconfck": "bin/tsconfck.js" } }, "sha512-ks6Vjr/jEw0P1gmOVwutM3B7fWxoWBL2KRDb1JfqGVawBmO5UsvmWOQFGHBPl5yxYz4eERr19E6L7NMv+Fej4w=="],
@@ -6590,6 +6625,8 @@

    "builder-util/js-yaml": ["js-yaml@4.3.0", "", { "dependencies": { "argparse": "^2.0.1" }, "bin": { "js-yaml": "bin/js-yaml.js" } }, "sha512-1td788aAnnZ5qs7V2QIRl1owjtYpbKt749Y3xauqQgwIIGF/xXWz1wMTEBx5O3LK3lXLVuqXPdPxj2BoFHaW9Q=="],

+    "cbor-extract/node-gyp-build-optional-packages": ["node-gyp-build-optional-packages@5.1.1", "", { "dependencies": { "detect-libc": "^2.0.1" }, "bin": { "node-gyp-build-optional-packages": "bin.js", "node-gyp-build-optional-packages-test": "build-test.js", "node-gyp-build-optional-packages-optional": "optional.js" } }, "sha512-+P72GAjVAbTxjjwUmwjVrqrdZROD4nf8KgpBoDxqXXTiYZZt/ud60dE5yvCSr9lRO8e8yv6kgJIC0K0PfZFVQw=="],
+
    "cliui/string-width": ["string-width@4.2.3", "", { "dependencies": { "emoji-regex": "^8.0.0", "is-fullwidth-code-point": "^3.0.0", "strip-ansi": "^6.0.1" } }, "sha512-wKyQRQpjJ0sIp62ErSZdGsjMJWsap5oRNihHhu6G7JVO/9jIB6UyevL+tXuOqrng8j/cxKTWyWUwvSTriiZz/g=="],

    "cliui/strip-ansi": ["strip-ansi@6.0.1", "", { "dependencies": { "ansi-regex": "^5.0.1" } }, "sha512-Y38VPSHcqkFrCpFnQ9vuSXmquuv5oXOKpGeT6aGrr3o3Gc9AlVa6JBfUSOCnbxGGZF+/0ooI7KrPuUSztUdU5A=="],
@@ -6734,6 +6771,8 @@

    "minipass-pipeline/minipass": ["minipass@3.3.6", "", { "dependencies": { "yallist": "^4.0.0" } }, "sha512-DxiNidxSEK+tHG6zOIklvNOwm3hvCrbUrdtzY74U6HKTJxvIDfOUL5W5P2Ghd3DTkhhKPYGqeNUIh5qcM4YBfw=="],

+    "modal/uuid": ["uuid@11.1.1", "", { "bin": { "uuid": "dist/esm/bin/uuid" } }, "sha512-vIYxrBCC/N/K+Js3qSN88go7kIfNPssr/hHCesKCQNAjmgvYS2oqr69kIufEG+O4+PfezOH4EbIeHCfFov8ZgQ=="],
+
    "motion/framer-motion": ["framer-motion@12.42.2", "", { "dependencies": { "motion-dom": "^12.42.2", "motion-utils": "^12.39.0", "tslib": "^2.4.0" }, "peerDependencies": { "@emotion/is-prop-valid": "*", "react": "^18.0.0 || ^19.0.0", "react-dom": "^18.0.0 || ^19.0.0" }, "optionalPeers": ["@emotion/is-prop-valid", "react", "react-dom"] }, "sha512-5XY9luDiu0oHfHBjpDthFMh0ES+122w6p/papSJBweMkO8Sn+PW2QaEgRblQBpWFnuvZS5qvarpt/hO2pjGmnw=="],

    "nitro/h3": ["h3@2.0.1-rc.5", "", { "dependencies": { "rou3": "^0.7.9", "srvx": "^0.9.1" }, "peerDependencies": { "crossws": "^0.4.1" }, "optionalPeers": ["crossws"] }, "sha512-qkohAzCab0nLzXNm78tBjZDvtKMTmtygS8BJLT3VPczAQofdqlFXDPkXdLMJN4r05+xqneG8snZJ0HgkERCZTg=="],
@@ -25,20 +25,8 @@ import { ToolSchemaProjection } from "./utils/tool-schema"

 const ADAPTER = "gemini"
 const MEDIA_MIMES = new Set<string>(ProviderShared.MEDIA_MIMES)
-// Google documents this sentinel for replaying Gemini 3 function calls after their original signature was lost.
-const SKIP_THOUGHT_SIGNATURE_VALIDATOR = "skip_thought_signature_validator"
 export const DEFAULT_BASE_URL = "https://generativelanguage.googleapis.com/v1beta"

-// Gemini 3 rejects replayed function calls without a thought signature. Google's SDKs avoid that in normal chats by
-// retaining complete model responses, but OpenCode reconstructs durable history and may encounter an unsigned call
-// from an older or external session. Model IDs are open-ended, so unknown Gemini aliases inherit current behavior.
-const requiresThoughtSignatureFallback = (modelID: string) => {
-  if (!/(^|\/)gemini-/i.test(modelID)) return false
-  if (/(^|\/)gemini-(?:1|2)(?:[.-]|$)/i.test(modelID)) return false
-  if (/(^|\/)gemini-pro(?:-vision)?$/i.test(modelID)) return false
-  return !/(^|\/)gemini-robotics-er-1\.5(?:[.-]|$)/i.test(modelID)
-}
-
 export interface OptionsInput {
  readonly [key: string]: unknown
  readonly cachedContent?: string
@@ -157,9 +145,6 @@ const GeminiGenerationConfig = Schema.Struct({
  temperature: Schema.optional(Schema.Number),
  topP: Schema.optional(Schema.Number),
  topK: Schema.optional(Schema.Number),
-  frequencyPenalty: Schema.optional(Schema.Number),
-  presencePenalty: Schema.optional(Schema.Number),
-  seed: Schema.optional(Schema.Number),
  stopSequences: optionalArray(Schema.String),
  thinkingConfig: Schema.optional(GeminiThinkingConfig),
 })
@@ -217,13 +202,11 @@ interface ParserState {
 //    keys on non-object scalars. Mirrors OpenCode's historical Gemini rules.
 //
 // 2. Project — lossy mapping from JSON Schema to Gemini's schema dialect:
-//    drop empty root parameter schemas while preserving nested empty objects,
-//    expand type arrays into `anyOf`, derive `nullable: true` from null members,
-//    coerce `const` to `[const]` enum, recurse properties/items, and propagate
+//    drop empty objects, derive `nullable: true` from `type: [..., "null"]`,
+//    coerce `const` to `[const]` enum, recurse properties/items, propagate
 //    only an allowlisted set of keys (description, required, format, type,
-//    nullable, enum, properties, items, allOf, anyOf, oneOf, minLength).
-//    Anything outside the allowlist (e.g. `additionalProperties`, `$ref`) is
-//    silently dropped.
+//    properties, items, allOf, anyOf, oneOf, minLength). Anything outside the
+//    allowlist (e.g. `additionalProperties`, `$ref`) is silently dropped.
 //
 // Sanitize runs first, then project. The implementation lives in
 // `utils/gemini-tool-schema` so this protocol keeps the same shape as the other
@@ -299,8 +282,6 @@ const lowerMessages = Effect.fn("Gemini.lowerMessages")(function* (request: LLMR

    if (message.role === "assistant") {
      const parts: Array<Schema.Schema.Type<typeof GeminiContentPart>> = []
-      // Parallel Gemini 3 calls may carry one signature on the first call; unsigned sibling calls are valid.
-      let hasSignedToolCall = false
      for (const part of message.content) {
        if (!ProviderShared.supportsContent(part, ["text", "reasoning", "tool-call"]))
          return yield* ProviderShared.unsupportedContent("Gemini", "assistant", ["text", "reasoning", "tool-call"])
@@ -313,17 +294,7 @@ const lowerMessages = Effect.fn("Gemini.lowerMessages")(function* (request: LLMR
          continue
        }
        if (part.type === "tool-call") {
-          const lowered = lowerToolCall(part)
-          const signature = lowered.thoughtSignature
-          parts.push({
-            ...lowered,
-            thoughtSignature:
-              signature ??
-              (requiresThoughtSignatureFallback(request.model.id) && !hasSignedToolCall
-                ? SKIP_THOUGHT_SIGNATURE_VALIDATOR
-                : undefined),
-          })
-          if (signature !== undefined) hasSignedToolCall = true
+          parts.push(lowerToolCall(part))
          continue
        }
      }
@@ -417,9 +388,6 @@ const fromRequest = Effect.fn("Gemini.fromRequest")(function* (request: LLMReque
    temperature: generation?.temperature,
    topP: generation?.topP,
    topK: generation?.topK,
-    frequencyPenalty: generation?.frequencyPenalty,
-    presencePenalty: generation?.presencePenalty,
-    seed: generation?.seed,
    stopSequences: generation?.stop,
    thinkingConfig: options.thinkingConfig,
  }
@@ -90,15 +90,10 @@ const OpenResponsesFunctionCallOutput = Schema.Union([
 ])

 export const InputItem = Schema.Union([
-  Schema.Struct({ role: Schema.tag("system"), id: Schema.optionalKey(Schema.String), content: Schema.String }),
-  Schema.Struct({
-    role: Schema.tag("user"),
-    id: Schema.optionalKey(Schema.String),
-    content: Schema.Array(OpenResponsesInputContent),
-  }),
+  Schema.Struct({ role: Schema.tag("system"), content: Schema.String }),
+  Schema.Struct({ role: Schema.tag("user"), content: Schema.Array(OpenResponsesInputContent) }),
  Schema.Struct({
    role: Schema.tag("assistant"),
-    id: Schema.optionalKey(Schema.String),
    content: Schema.Array(OpenResponsesOutputText),
    phase: Schema.optionalKey(MessagePhase),
  }),
@@ -106,23 +101,19 @@ export const InputItem = Schema.Union([
  OpenResponsesItemReference,
  Schema.Struct({
    type: Schema.tag("function_call"),
-    id: Schema.optionalKey(Schema.String),
    call_id: Schema.String,
    name: Schema.String,
    arguments: Schema.String,
  }),
  Schema.Struct({
    type: Schema.tag("function_call_output"),
-    id: Schema.optionalKey(Schema.String),
    call_id: Schema.String,
    output: OpenResponsesFunctionCallOutput,
  }),
 ])
 type OpenResponsesInputItem = Schema.Schema.Type<typeof InputItem>
-type ProviderInputItem = Readonly<Record<string, unknown>> & { readonly type: string; readonly id?: string }
 type LoweredInputItem =
  | OpenResponsesInputItem
-  | ProviderInputItem
  | {
      readonly role: "assistant"
      readonly content: ReadonlyArray<{ readonly type: "output_text"; readonly text: string }>
@@ -137,7 +128,7 @@ type OpenResponsesReasoningInput = {
  summary: Array<{ type: "summary_text"; text: string }>
  encrypted_content?: string | null
 }
-type OpenResponsesReasoningReplay = Omit<OpenResponsesReasoningInput, "id"> & { id?: string }
+type OpenResponsesReasoningReplay = Omit<OpenResponsesReasoningInput, "id">

 export const Tool = Schema.Struct({
  type: Schema.tag("function"),
@@ -263,11 +254,6 @@ export interface Extension {
    readonly request: LLMRequest
  }) => MediaInput | undefined
  readonly messagePhase?: (value: unknown) => MessagePhase | null | undefined
-  readonly lowerProviderItem?: (
-    part: ToolResultPart,
-    providerMetadataKey: string,
-    store: boolean | undefined,
-  ) => ProviderInputItem | undefined
 }

 const BASE: Extension = { id: ADAPTER, name: NAME }
@@ -324,17 +310,6 @@ export const lowerToolChoice = (protocolName: string, toolChoice: NonNullable<LL
    tool: (toolName) => ({ type: "function" as const, name: toolName }),
  })

-const metadataItemID = (
-  part: { readonly itemId?: string; readonly providerMetadata?: ProviderMetadata },
-  providerMetadataKey: string,
-) => {
-  if (part.itemId) return part.itemId
-  const metadata = part.providerMetadata?.[providerMetadataKey]
-  return ProviderShared.isRecord(metadata) && typeof metadata.itemId === "string" && metadata.itemId.length > 0
-    ? metadata.itemId
-    : undefined
-}
-
 const lowerToolCall = (part: ToolCallPart): OpenResponsesInputItem => ({
  type: "function_call",
  call_id: part.id,
@@ -344,23 +319,26 @@ const lowerToolCall = (part: ToolCallPart): OpenResponsesInputItem => ({

 const lowerReasoning = (part: ReasoningPart, providerMetadataKey: string): OpenResponsesReasoningInput | undefined => {
  const metadata = part.providerMetadata?.[providerMetadataKey]
-  const itemId = metadataItemID(part, providerMetadataKey)
-  if (!itemId) return undefined
+  if (!ProviderShared.isRecord(metadata) || typeof metadata.itemId !== "string" || metadata.itemId.length === 0)
+    return undefined
  const encryptedContent =
-    ProviderShared.isRecord(metadata) &&
-    (typeof metadata.reasoningEncryptedContent === "string" || metadata.reasoningEncryptedContent === null)
+    typeof metadata.reasoningEncryptedContent === "string" || metadata.reasoningEncryptedContent === null
      ? metadata.reasoningEncryptedContent
      : undefined
  return {
    type: "reasoning",
-    id: itemId,
+    id: metadata.itemId,
    summary: part.text.length > 0 ? [{ type: "summary_text", text: part.text }] : [],
    encrypted_content: encryptedContent,
  }
 }

-const hostedToolItemID = (part: ToolResultPart, providerMetadataKey: string) =>
-  metadataItemID(part, providerMetadataKey)
+const hostedToolItemID = (part: ToolResultPart, providerMetadataKey: string) => {
+  const metadata = part.providerMetadata?.[providerMetadataKey]
+  return ProviderShared.isRecord(metadata) && typeof metadata.itemId === "string" && metadata.itemId.length > 0
+    ? metadata.itemId
+    : undefined
+}

 const lowerMedia = Effect.fn("OpenResponses.lowerMedia")(function* (
  part: MediaPart,
@@ -419,18 +397,17 @@ const lowerToolResultOutput = Effect.fn("OpenResponses.lowerToolResultOutput")(f
 })

 const lowerMessages = Effect.fn("OpenResponses.lowerMessages")(function* (request: LLMRequest, extension: Extension) {
-  const options = OpenResponsesOptions.resolve(request)
  const system: LoweredInputItem[] =
    request.system.length === 0 ? [] : [{ role: "system", content: ProviderShared.joinText(request.system) }]
  const input: LoweredInputItem[] = [...system]
-  const store = options.store
+  const store = OpenResponsesOptions.resolve(request).store
  const providerMetadataKey = request.model.route.providerMetadataKey ?? "openresponses"

  for (const message of request.messages) {
    if (message.role === "system") {
      const part = yield* ProviderShared.wrappedSystemUpdate(extension.name, message)
      const previous = input.at(-1)
-      if (previous && "role" in previous && previous.role === "user" && Array.isArray(previous.content))
+      if (previous && "role" in previous && previous.role === "user")
        input[input.length - 1] = {
          role: "user",
          content: [...previous.content, { type: "input_text", text: part.text }],
@@ -450,24 +427,24 @@ const lowerMessages = Effect.fn("OpenResponses.lowerMessages")(function* (reques
    if (message.role === "assistant") {
      const content: TextPart[] = []
      const reasoningItems: Record<string, OpenResponsesReasoningReplay> = {}
-      const hostedToolItems = new Set<string>()
+      const reasoningReferences = new Set<string>()
+      const hostedToolReferences = new Set<string>()
      const flushText = () => {
        if (content.length === 0) return
-        const groups = content.reduce<
-          Array<{ phase: MessagePhase | null | undefined; itemId: string | undefined; parts: TextPart[] }>
-        >((groups, part) => {
-          const metadata = part.providerMetadata?.[providerMetadataKey]
-          const phase = ProviderShared.isRecord(metadata) ? messagePhase(metadata.phase, extension) : undefined
-          const itemId = metadataItemID(part, providerMetadataKey)
-          const group = groups.at(-1)
-          if (group && group.phase === phase && group.itemId === itemId) group.parts.push(part)
-          else groups.push({ phase, itemId, parts: [part] })
-          return groups
-        }, [])
+        const groups = content.reduce<Array<{ phase: MessagePhase | null | undefined; parts: TextPart[] }>>(
+          (groups, part) => {
+            const metadata = part.providerMetadata?.[providerMetadataKey]
+            const phase = ProviderShared.isRecord(metadata) ? messagePhase(metadata.phase, extension) : undefined
+            const group = groups.at(-1)
+            if (group && group.phase === phase) group.parts.push(part)
+            else groups.push({ phase, parts: [part] })
+            return groups
+          },
+          [],
+        )
        input.push(
          ...groups.map((group) => ({
            role: "assistant" as const,
-            ...(group.itemId === undefined ? {} : { id: group.itemId }),
            content: group.parts.map((part) => ({ type: "output_text" as const, text: part.text })),
            ...(group.phase === undefined ? {} : { phase: group.phase }),
          })),
@@ -483,6 +460,11 @@ const lowerMessages = Effect.fn("OpenResponses.lowerMessages")(function* (reques
          flushText()
          const reasoning = lowerReasoning(part, providerMetadataKey)
          if (!reasoning) continue
+          if (store !== false) {
+            if (!reasoningReferences.has(reasoning.id)) input.push({ type: "item_reference", id: reasoning.id })
+            reasoningReferences.add(reasoning.id)
+            continue
+          }
          const existing = reasoningItems[reasoning.id]
          if (existing) {
            existing.summary.push(...reasoning.summary)
@@ -492,7 +474,6 @@ const lowerMessages = Effect.fn("OpenResponses.lowerMessages")(function* (reques
          }
          const replay = {
            type: reasoning.type,
-            id: reasoning.id,
            summary: reasoning.summary,
            encrypted_content: reasoning.encrypted_content,
          }
@@ -509,18 +490,16 @@ const lowerMessages = Effect.fn("OpenResponses.lowerMessages")(function* (reques
        if (part.type === "tool-result" && part.providerExecuted === true) {
          flushText()
          const itemID = hostedToolItemID(part, providerMetadataKey)
-          const providerItem = extension.lowerProviderItem?.(part, providerMetadataKey, store)
-          if (providerItem && itemID && !hostedToolItems.has(itemID)) input.push(providerItem)
-          if (!providerItem && store !== false && itemID && !hostedToolItems.has(itemID))
+          if (store !== false && itemID && !hostedToolReferences.has(itemID))
            input.push({ type: "item_reference", id: itemID })
-          if (!providerItem && store === false && part.result.type === "content") {
+          if (store === false && part.result.type === "content") {
            const content: ReadonlyArray<Content> = part.result.value
            input.push({
              role: "user",
              content: yield* Effect.forEach(content, (item) => lowerToolResultContentItem(item, request, extension)),
            })
          }
-          if (itemID) hostedToolItems.add(itemID)
+          if (itemID) hostedToolReferences.add(itemID)
          continue
        }
        return yield* ProviderShared.unsupportedContent(extension.name, "assistant", [
@@ -662,9 +641,9 @@ const onOutputTextDelta = (state: ParserState, event: Event, id: string): StepRe
  if (!event.delta) return [state, NO_EVENTS]
  const events: LLMEvent[] = []
  const phase = state.messagePhases[id]
-  const metadata = providerMetadata(state, { itemId: id, ...(phase === undefined ? {} : { phase }) })
-  const lifecycle = Lifecycle.textStart(state.lifecycle, events, id, metadata, id)
-  return [{ ...state, lifecycle: Lifecycle.textDelta(lifecycle, events, id, event.delta, metadata, id) }, events]
+  const metadata = phase === undefined ? undefined : providerMetadata(state, { phase })
+  const lifecycle = Lifecycle.textStart(state.lifecycle, events, id, metadata)
+  return [{ ...state, lifecycle: Lifecycle.textDelta(lifecycle, events, id, event.delta) }, events]
 }

 const onOutputTextDone = (state: ParserState, event: Event, id: string): StepResult => {
@@ -673,13 +652,7 @@ const onOutputTextDone = (state: ParserState, event: Event, id: string): StepRes
    return onOutputTextDelta(state, { ...event, delta: event.text }, id)
  }
  const events: LLMEvent[] = []
-  return [
-    {
-      ...state,
-      lifecycle: Lifecycle.textEnd(state.lifecycle, events, id, providerMetadata(state, { itemId: id }), id),
-    },
-    events,
-  ]
+  return [{ ...state, lifecycle: Lifecycle.textEnd(state.lifecycle, events, id) }, events]
 }

 export const onReasoningDelta = (state: ParserState, event: Event, itemID: string): StepResult => {
@@ -690,14 +663,7 @@ export const onReasoningDelta = (state: ParserState, event: Event, itemID: strin
  return [
    {
      ...state,
-      lifecycle: Lifecycle.reasoningDelta(
-        state.lifecycle,
-        events,
-        id,
-        event.delta,
-        providerMetadata(state, { itemId: itemID }),
-        itemID,
-      ),
+      lifecycle: Lifecycle.reasoningDelta(state.lifecycle, events, id, event.delta),
    },
    events,
  ]
@@ -739,13 +705,7 @@ const onOutputItemAdded = (state: ParserState, event: Event): StepResult => {
    return [
      {
        ...state,
-        lifecycle: Lifecycle.reasoningStart(
-          state.lifecycle,
-          events,
-          `${item.id}:0`,
-          reasoningMetadata(state, item),
-          item.id,
-        ),
+        lifecycle: Lifecycle.reasoningStart(state.lifecycle, events, `${item.id}:0`, reasoningMetadata(state, item)),
        reasoningItems: {
          ...state.reasoningItems,
          [item.id]: { encryptedContent: item.encrypted_content, summaryParts: { 0: "active" } },
@@ -764,7 +724,6 @@ const onOutputItemAdded = (state: ParserState, event: Event): StepResult => {
      lifecycle,
      tools: ToolStream.start(state.tools, item.id, {
        id: item.call_id ?? item.id,
-        itemId: item.id,
        name: item.name ?? "",
        input: item.arguments ?? "",
        providerMetadata: metadata,
@@ -772,12 +731,7 @@ const onOutputItemAdded = (state: ParserState, event: Event): StepResult => {
    },
    [
      ...events,
-      LLMEvent.toolInputStart({
-        id: item.call_id ?? item.id,
-        itemId: item.id,
-        name: item.name ?? "",
-        providerMetadata: metadata,
-      }),
+      LLMEvent.toolInputStart({ id: item.call_id ?? item.id, name: item.name ?? "", providerMetadata: metadata }),
    ],
  ]
 }
@@ -796,7 +750,6 @@ const onReasoningSummaryPartAdded = (state: ParserState, event: Event): StepResu
          events,
          `${event.item_id}:0`,
          providerMetadata(state, { itemId: event.item_id, reasoningEncryptedContent: null }),
-          event.item_id,
        ),
        reasoningItems: {
          ...state.reasoningItems,
@@ -817,7 +770,6 @@ const onReasoningSummaryPartAdded = (state: ParserState, event: Event): StepResu
          events,
          `${event.item_id}:${entry[0]}`,
          providerMetadata(state, { itemId: event.item_id }),
-          event.item_id,
        ),
      state.lifecycle,
    )
@@ -829,7 +781,6 @@ const onReasoningSummaryPartAdded = (state: ParserState, event: Event): StepResu
        events,
        `${event.item_id}:${event.summary_index}`,
        providerMetadata(state, { itemId: event.item_id, reasoningEncryptedContent: item.encryptedContent ?? null }),
-        event.item_id,
      ),
      reasoningItems: {
        ...state.reasoningItems,
@@ -865,7 +816,6 @@ const onReasoningSummaryPartDone = (state: ParserState, event: Event): StepResul
              events,
              `${event.item_id}:${event.summary_index}`,
              providerMetadata(state, { itemId: event.item_id }),
-              event.item_id,
            )
          : state.lifecycle,
      reasoningItems: {
@@ -920,8 +870,7 @@ const onOutputItemDone = Effect.fn("OpenResponses.onOutputItemDone")(function* (
          state.lifecycle,
          events,
          item.id,
-          providerMetadata(state, { itemId: item.id, ...(phase === undefined ? {} : { phase }) }),
-          item.id,
+          phase === undefined ? undefined : providerMetadata(state, { phase }),
        ),
        messageItems,
        messagePhases,
@@ -932,15 +881,9 @@ const onOutputItemDone = Effect.fn("OpenResponses.onOutputItemDone")(function* (

  if (item.type === "function_call") {
    if (!item.id || !item.call_id || !item.name) return [state, NO_EVENTS] satisfies StepResult
-    const metadata = providerMetadata(state, { itemId: item.id })
    const tools = state.tools[item.id]
      ? state.tools
-      : ToolStream.start(state.tools, item.id, {
-          id: item.call_id,
-          itemId: item.id,
-          name: item.name,
-          providerMetadata: metadata,
-        })
+      : ToolStream.start(state.tools, item.id, { id: item.call_id, name: item.name })
    const result =
      item.arguments === undefined
        ? yield* ToolStream.finish(state.id, tools, item.id)
@@ -970,7 +913,7 @@ const onOutputItemDone = Effect.fn("OpenResponses.onOutputItemDone")(function* (
      const lifecycle = Object.entries(reasoningItem.summaryParts)
        .filter((entry) => entry[1] === "active" || entry[1] === "can-conclude")
        .reduce(
-          (lifecycle, entry) => Lifecycle.reasoningEnd(lifecycle, events, `${item.id}:${entry[0]}`, metadata, item.id),
+          (lifecycle, entry) => Lifecycle.reasoningEnd(lifecycle, events, `${item.id}:${entry[0]}`, metadata),
          state.lifecycle,
        )
      const { [item.id]: _removed, ...reasoningItems } = state.reasoningItems
@@ -978,12 +921,12 @@ const onOutputItemDone = Effect.fn("OpenResponses.onOutputItemDone")(function* (
    }
    if (!state.lifecycle.reasoning.has(item.id)) {
      const lifecycle = Lifecycle.stepStart(state.lifecycle, events)
-      events.push(LLMEvent.reasoningStart({ id: item.id, itemId: item.id, providerMetadata: metadata }))
-      events.push(LLMEvent.reasoningEnd({ id: item.id, itemId: item.id, providerMetadata: metadata }))
+      events.push(LLMEvent.reasoningStart({ id: item.id, providerMetadata: metadata }))
+      events.push(LLMEvent.reasoningEnd({ id: item.id, providerMetadata: metadata }))
      return [{ ...state, lifecycle }, events] satisfies StepResult
    }
    return [
-      { ...state, lifecycle: Lifecycle.reasoningEnd(state.lifecycle, events, item.id, metadata, item.id) },
+      { ...state, lifecycle: Lifecycle.reasoningEnd(state.lifecycle, events, item.id, metadata) },
      events,
    ] satisfies StepResult
  }
@@ -38,14 +38,10 @@ const OpenAIResponsesToolChoice = Schema.Union([
 const OpenAIResponsesInputItem = Schema.Union([
  Schema.Struct({
    role: Schema.tag("assistant"),
-    id: Schema.optionalKey(Schema.String),
    content: Schema.Array(Schema.Struct({ type: Schema.tag("output_text"), text: Schema.String })),
    phase: Schema.optionalKey(Schema.NullOr(OpenResponses.MessagePhase)),
  }),
  OpenResponses.InputItem,
-  Schema.StructWithRest(Schema.Struct({ type: Schema.String, id: Schema.optionalKey(Schema.String) }), [
-    Schema.Record(Schema.String, Schema.Unknown),
-  ]),
 ])

 const OpenAIResponsesCoreFields = {
@@ -84,25 +80,6 @@ const extension = {
      mime_type: media.mime,
    }
  },
-  lowerProviderItem: (part, providerMetadataKey, store) => {
-    const metadata = part.providerMetadata?.[providerMetadataKey]
-    if (!ProviderShared.isRecord(metadata) || !ProviderShared.isRecord(metadata.item)) return undefined
-    if (typeof metadata.item.type !== "string") return undefined
-    const id = typeof metadata.item.id === "string" ? metadata.item.id : undefined
-    // The public API requires stored state to replay image-generation items. In
-    // stateless mode, lower the generated file through the existing user-image fallback.
-    if (metadata.item.type === "image_generation_call" && store === false) return undefined
-    if (metadata.item.type === "image_generation_call")
-      return {
-        type: metadata.item.type,
-        ...(id === undefined ? {} : { id }),
-        ...(typeof metadata.item.status === "string" ? { status: metadata.item.status } : {}),
-        ...(typeof metadata.item.revised_prompt === "string" ? { revised_prompt: metadata.item.revised_prompt } : {}),
-        ...(typeof metadata.item.result === "string" ? { result: metadata.item.result } : {}),
-      }
-    const item: Record<string, unknown> & { type: string } = { ...metadata.item, type: metadata.item.type }
-    return item
-  },
 } satisfies OpenResponses.Extension

 const nativeImageToolInput = (tool: ToolDefinition) => {
@@ -218,29 +195,23 @@ const onHostedToolDone = Effect.fn("OpenAIResponses.onHostedToolDone")(function*
  item: HostedToolItem,
 ) {
  const tool = HOSTED_TOOLS[item.type]
-  const callMetadata = OpenResponses.providerMetadata(state, { itemId: item.id })
-  const resultMetadata = OpenResponses.providerMetadata(
-    state,
-    item.type === "image_generation_call" ? { itemId: item.id } : { itemId: item.id, item },
-  )
+  const providerMetadata = OpenResponses.providerMetadata(state, { itemId: item.id })
  const events: LLMEvent[] = []
  const lifecycle = Lifecycle.stepStart(state.lifecycle, events)
  events.push(
    LLMEvent.toolCall({
      id: item.id,
-      itemId: item.id,
      name: tool.name,
      input: tool.input(item),
      providerExecuted: true,
-      providerMetadata: callMetadata,
+      providerMetadata,
    }),
    LLMEvent.toolResult({
      id: item.id,
-      itemId: item.id,
      name: tool.name,
      result: yield* hostedToolResult(item),
      providerExecuted: true,
-      providerMetadata: resultMetadata,
+      providerMetadata,
    }),
  )
  return [{ ...state, lifecycle }, events] satisfies OpenResponses.StepResult
@@ -61,57 +61,37 @@ const emptyObjectSchema = (schema: Record<string, unknown>) =>
  (!isRecord(schema.properties) || Object.keys(schema.properties).length === 0) &&
  !schema.additionalProperties

-const projectNode = (schema: unknown, nested = false): Record<string, unknown> | undefined => {
+const projectNode = (schema: unknown): Record<string, unknown> | undefined => {
  if (!isRecord(schema)) return undefined
-  if (!nested && emptyObjectSchema(schema)) return undefined
-  const types = Array.isArray(schema.type) ? schema.type.filter((type) => type !== "null") : undefined
-  const anyOf = Array.isArray(schema.anyOf) ? schema.anyOf : undefined
-  const hasNullAnyOf = anyOf?.some((item) => isRecord(item) && item.type === "null") ?? false
-  const anyOfTypes = hasNullAnyOf ? anyOf?.filter((item) => !isRecord(item) || item.type !== "null") : anyOf
-  const flattenedAnyOf = hasNullAnyOf && anyOfTypes?.length === 1 ? projectNode(anyOfTypes[0], true) : undefined
-  const result = Object.fromEntries(
+  if (emptyObjectSchema(schema)) return undefined
+  return Object.fromEntries(
    [
      ["description", schema.description],
      ["required", schema.required],
      ["format", schema.format],
-      ["type", types ? (types.length === 0 ? "null" : undefined) : schema.type],
-      [
-        "nullable",
-        (Array.isArray(schema.type) && schema.type.includes("null") && types && types.length > 0) || hasNullAnyOf
-          ? true
-          : undefined,
-      ],
+      ["type", Array.isArray(schema.type) ? schema.type.filter((type) => type !== "null")[0] : schema.type],
+      ["nullable", Array.isArray(schema.type) && schema.type.includes("null") ? true : undefined],
      ["enum", schema.const !== undefined ? [schema.const] : schema.enum],
      [
        "properties",
        isRecord(schema.properties)
-          ? Object.fromEntries(Object.entries(schema.properties).map(([key, value]) => [key, projectNode(value, true)]))
+          ? Object.fromEntries(Object.entries(schema.properties).map(([key, value]) => [key, projectNode(value)]))
          : undefined,
      ],
      [
        "items",
        Array.isArray(schema.items)
-          ? schema.items.map((item) => projectNode(item, true))
+          ? schema.items.map(projectNode)
          : schema.items === undefined
            ? undefined
-            : projectNode(schema.items, true),
+            : projectNode(schema.items),
      ],
-      ["allOf", Array.isArray(schema.allOf) ? schema.allOf.map((item) => projectNode(item, true)) : undefined],
-      [
-        "anyOf",
-        anyOfTypes
-          ? hasNullAnyOf && anyOfTypes.length === 1
-            ? undefined
-            : anyOfTypes.map((item) => projectNode(item, true))
-          : types && types.length > 0
-            ? types.map((type) => ({ type }))
-            : undefined,
-      ],
-      ["oneOf", Array.isArray(schema.oneOf) ? schema.oneOf.map((item) => projectNode(item, true)) : undefined],
+      ["allOf", Array.isArray(schema.allOf) ? schema.allOf.map(projectNode) : undefined],
+      ["anyOf", Array.isArray(schema.anyOf) ? schema.anyOf.map(projectNode) : undefined],
+      ["oneOf", Array.isArray(schema.oneOf) ? schema.oneOf.map(projectNode) : undefined],
      ["minLength", schema.minLength],
    ].filter((entry) => entry[1] !== undefined),
  )
-  return flattenedAnyOf ? { ...result, ...flattenedAnyOf } : result
 }

 export const convert = (schema: unknown) => projectNode(sanitizeNode(schema))
@@ -1,10 +1,4 @@
-import {
-  LLMEvent,
-  type FinishReasonDetails,
-  type ProviderMetadata,
-  type ResponseItemID,
-  type Usage,
-} from "../../schema"
+import { LLMEvent, type FinishReasonDetails, type ProviderMetadata, type Usage } from "../../schema"

 export interface State {
  readonly stepStarted: boolean
@@ -20,29 +14,16 @@ export const stepStart = (state: State, events: LLMEvent[]): State => {
  return { ...state, stepStarted: true }
 }

-export const textStart = (
-  state: State,
-  events: LLMEvent[],
-  id: string,
-  providerMetadata?: ProviderMetadata,
-  itemId?: ResponseItemID,
-): State => {
+export const textStart = (state: State, events: LLMEvent[], id: string, providerMetadata?: ProviderMetadata): State => {
  if (state.text.has(id)) return state
  const stepped = stepStart(state, events)
-  events.push(LLMEvent.textStart({ id, ...(itemId === undefined ? {} : { itemId }), providerMetadata }))
+  events.push(LLMEvent.textStart({ id, providerMetadata }))
  return { ...stepped, text: new Set([...stepped.text, id]) }
 }

-export const textDelta = (
-  state: State,
-  events: LLMEvent[],
-  id: string,
-  text: string,
-  providerMetadata?: ProviderMetadata,
-  itemId?: ResponseItemID,
-): State => {
-  const started = textStart(state, events, id, providerMetadata, itemId)
-  events.push(LLMEvent.textDelta({ id, ...(itemId === undefined ? {} : { itemId }), text, providerMetadata }))
+export const textDelta = (state: State, events: LLMEvent[], id: string, text: string): State => {
+  const started = textStart(state, events, id)
+  events.push(LLMEvent.textDelta({ id, text }))
  return started
 }

@@ -51,11 +32,10 @@ export const reasoningStart = (
  events: LLMEvent[],
  id: string,
  providerMetadata?: ProviderMetadata,
-  itemId?: ResponseItemID,
 ): State => {
  if (state.reasoning.has(id)) return state
  const stepped = stepStart(state, events)
-  events.push(LLMEvent.reasoningStart({ id, ...(itemId === undefined ? {} : { itemId }), providerMetadata }))
+  events.push(LLMEvent.reasoningStart({ id, providerMetadata }))
  return { ...stepped, reasoning: new Set([...stepped.reasoning, id]) }
 }

@@ -65,10 +45,9 @@ export const reasoningDelta = (
  id: string,
  text: string,
  providerMetadata?: ProviderMetadata,
-  itemId?: ResponseItemID,
 ): State => {
-  const started = reasoningStart(state, events, id, providerMetadata, itemId)
-  events.push(LLMEvent.reasoningDelta({ id, ...(itemId === undefined ? {} : { itemId }), text, providerMetadata }))
+  const started = reasoningStart(state, events, id, providerMetadata)
+  events.push(LLMEvent.reasoningDelta({ id, text, providerMetadata }))
  return started
 }

@@ -77,26 +56,19 @@ export const reasoningEnd = (
  events: LLMEvent[],
  id: string,
  providerMetadata?: ProviderMetadata,
-  itemId?: ResponseItemID,
 ): State => {
  if (!state.reasoning.has(id)) return state
  const stepped = stepStart(state, events)
-  events.push(LLMEvent.reasoningEnd({ id, ...(itemId === undefined ? {} : { itemId }), providerMetadata }))
+  events.push(LLMEvent.reasoningEnd({ id, providerMetadata }))
  const reasoning = new Set(stepped.reasoning)
  reasoning.delete(id)
  return { ...stepped, reasoning }
 }

-export const textEnd = (
-  state: State,
-  events: LLMEvent[],
-  id: string,
-  providerMetadata?: ProviderMetadata,
-  itemId?: ResponseItemID,
-): State => {
+export const textEnd = (state: State, events: LLMEvent[], id: string, providerMetadata?: ProviderMetadata): State => {
  if (!state.text.has(id)) return state
  const stepped = stepStart(state, events)
-  events.push(LLMEvent.textEnd({ id, ...(itemId === undefined ? {} : { itemId }), providerMetadata }))
+  events.push(LLMEvent.textEnd({ id, providerMetadata }))
  const text = new Set(stepped.text)
  text.delete(id)
  return { ...stepped, text }
@@ -1,12 +1,5 @@
 import { Effect } from "effect"
-import {
-  AIError,
-  LLMEvent,
-  type ProviderMetadata,
-  type ResponseItemID,
-  type ToolCall,
-  type ToolInputError,
-} from "../../schema"
+import { AIError, LLMEvent, type ProviderMetadata, type ToolCall, type ToolInputError } from "../../schema"
 import { eventError, parseToolInput, type ToolAccumulator } from "../shared"

 type StreamKey = string | number
@@ -17,7 +10,6 @@ type StreamKey = string | number
 * so far, not the parsed object.
 */
 export interface PendingTool extends ToolAccumulator {
-  readonly itemId?: ResponseItemID
  readonly providerExecuted?: boolean
  readonly providerMetadata?: ProviderMetadata
 }
@@ -60,7 +52,6 @@ const withoutTool = <K extends StreamKey>(tools: State<K>, key: K): State<K> =>
 const inputStart = (tool: PendingTool) =>
  LLMEvent.toolInputStart({
    id: tool.id,
-    ...(tool.itemId === undefined ? {} : { itemId: tool.itemId }),
    name: tool.name,
    providerExecuted: tool.providerExecuted ? true : undefined,
    providerMetadata: tool.providerMetadata,
@@ -69,7 +60,6 @@ const inputStart = (tool: PendingTool) =>
 const inputDelta = (tool: PendingTool, text: string) =>
  LLMEvent.toolInputDelta({
    id: tool.id,
-    ...(tool.itemId === undefined ? {} : { itemId: tool.itemId }),
    name: tool.name,
    text,
  })
@@ -80,7 +70,6 @@ const toolCall = (route: string, tool: PendingTool, inputOverride?: string) => {
    Effect.map((input): ToolCall | ToolInputError =>
      LLMEvent.toolCall({
        id: tool.id,
-        ...(tool.itemId === undefined ? {} : { itemId: tool.itemId }),
        name: tool.name,
        input,
        providerExecuted: tool.providerExecuted ? true : undefined,
@@ -93,7 +82,6 @@ const toolCall = (route: string, tool: PendingTool, inputOverride?: string) => {
        : Effect.succeed(
            LLMEvent.toolInputError({
              id: tool.id,
-              ...(tool.itemId === undefined ? {} : { itemId: tool.itemId }),
              name: tool.name,
              raw,
            }),
@@ -105,15 +93,7 @@ const toolCall = (route: string, tool: PendingTool, inputOverride?: string) => {
 const finishEvents = (tool: PendingTool, event: ToolCall | ToolInputError): ReadonlyArray<LLMEvent> =>
  event.type === "tool-input-error"
    ? [event]
-    : [
-        LLMEvent.toolInputEnd({
-          id: tool.id,
-          ...(tool.itemId === undefined ? {} : { itemId: tool.itemId }),
-          name: tool.name,
-          providerMetadata: tool.providerMetadata,
-        }),
-        event,
-      ]
+    : [LLMEvent.toolInputEnd({ id: tool.id, name: tool.name, providerMetadata: tool.providerMetadata }), event]

 /** Store the updated tool and produce the optional public delta event. */
 const appendTool = <K extends StreamKey>(
@@ -168,7 +148,6 @@ export const appendOrStart = <K extends StreamKey>(
    id,
    name,
    input: `${current?.input ?? ""}${delta.text}`,
-    itemId: current?.itemId,
    providerExecuted: current?.providerExecuted,
    providerMetadata: current?.providerMetadata,
  }
@@ -1,5 +1,5 @@
 import { Schema } from "effect"
-import { ContentBlockID, FinishReason, ProviderMetadata, ResponseItemID, ToolCallID } from "./ids"
+import { ContentBlockID, FinishReason, ProviderMetadata, ToolCallID } from "./ids"
 import { Message, ToolCallPart, ToolOutput, ToolResultPart, ToolResultValue, type ContentPart } from "./messages"
 import { ProviderFailureClassification } from "./errors"

@@ -84,7 +84,6 @@ export type StepStart = Schema.Schema.Type<typeof StepStart>
 export const TextStart = Schema.Struct({
  type: Schema.tag("text-start"),
  id: ContentBlockID,
-  itemId: Schema.optional(ResponseItemID),
  providerMetadata: Schema.optional(ProviderMetadata),
 }).annotate({ identifier: "LLM.Event.TextStart" })
 export type TextStart = Schema.Schema.Type<typeof TextStart>
@@ -93,7 +92,6 @@ export const TextDelta = Schema.Struct({
  type: Schema.tag("text-delta"),
  id: ContentBlockID,
  text: Schema.String,
-  itemId: Schema.optional(ResponseItemID),
  providerMetadata: Schema.optional(ProviderMetadata),
 }).annotate({ identifier: "LLM.Event.TextDelta" })
 export type TextDelta = Schema.Schema.Type<typeof TextDelta>
@@ -101,7 +99,6 @@ export type TextDelta = Schema.Schema.Type<typeof TextDelta>
 export const TextEnd = Schema.Struct({
  type: Schema.tag("text-end"),
  id: ContentBlockID,
-  itemId: Schema.optional(ResponseItemID),
  providerMetadata: Schema.optional(ProviderMetadata),
 }).annotate({ identifier: "LLM.Event.TextEnd" })
 export type TextEnd = Schema.Schema.Type<typeof TextEnd>
@@ -109,7 +106,6 @@ export type TextEnd = Schema.Schema.Type<typeof TextEnd>
 export const ReasoningStart = Schema.Struct({
  type: Schema.tag("reasoning-start"),
  id: ContentBlockID,
-  itemId: Schema.optional(ResponseItemID),
  providerMetadata: Schema.optional(ProviderMetadata),
 }).annotate({ identifier: "LLM.Event.ReasoningStart" })
 export type ReasoningStart = Schema.Schema.Type<typeof ReasoningStart>
@@ -118,7 +114,6 @@ export const ReasoningDelta = Schema.Struct({
  type: Schema.tag("reasoning-delta"),
  id: ContentBlockID,
  text: Schema.String,
-  itemId: Schema.optional(ResponseItemID),
  providerMetadata: Schema.optional(ProviderMetadata),
 }).annotate({ identifier: "LLM.Event.ReasoningDelta" })
 export type ReasoningDelta = Schema.Schema.Type<typeof ReasoningDelta>
@@ -126,7 +121,6 @@ export type ReasoningDelta = Schema.Schema.Type<typeof ReasoningDelta>
 export const ReasoningEnd = Schema.Struct({
  type: Schema.tag("reasoning-end"),
  id: ContentBlockID,
-  itemId: Schema.optional(ResponseItemID),
  providerMetadata: Schema.optional(ProviderMetadata),
 }).annotate({ identifier: "LLM.Event.ReasoningEnd" })
 export type ReasoningEnd = Schema.Schema.Type<typeof ReasoningEnd>
@@ -135,7 +129,6 @@ export const ToolInputStart = Schema.Struct({
  type: Schema.tag("tool-input-start"),
  id: ToolCallID,
  name: Schema.String,
-  itemId: Schema.optional(ResponseItemID),
  providerExecuted: Schema.optional(Schema.Boolean),
  providerMetadata: Schema.optional(ProviderMetadata),
 }).annotate({ identifier: "LLM.Event.ToolInputStart" })
@@ -144,7 +137,6 @@ export type ToolInputStart = Schema.Schema.Type<typeof ToolInputStart>
 export const ToolInputDelta = Schema.Struct({
  type: Schema.tag("tool-input-delta"),
  id: ToolCallID,
-  itemId: Schema.optional(ResponseItemID),
  name: Schema.String,
  text: Schema.String,
 }).annotate({ identifier: "LLM.Event.ToolInputDelta" })
@@ -154,7 +146,6 @@ export const ToolInputEnd = Schema.Struct({
  type: Schema.tag("tool-input-end"),
  id: ToolCallID,
  name: Schema.String,
-  itemId: Schema.optional(ResponseItemID),
  providerMetadata: Schema.optional(ProviderMetadata),
 }).annotate({ identifier: "LLM.Event.ToolInputEnd" })
 export type ToolInputEnd = Schema.Schema.Type<typeof ToolInputEnd>
@@ -163,7 +154,6 @@ export type ToolInputEnd = Schema.Schema.Type<typeof ToolInputEnd>
 export const ToolInputError = Schema.Struct({
  type: Schema.tag("tool-input-error"),
  id: ToolCallID,
-  itemId: Schema.optional(ResponseItemID),
  name: Schema.String,
  raw: Schema.String,
 }).annotate({ identifier: "LLM.Event.ToolInputError" })
@@ -172,7 +162,6 @@ export type ToolInputError = Schema.Schema.Type<typeof ToolInputError>
 export const ToolCall = Schema.Struct({
  type: Schema.tag("tool-call"),
  id: ToolCallID,
-  itemId: Schema.optional(ResponseItemID),
  name: Schema.String,
  input: Schema.Unknown,
  providerExecuted: Schema.optional(Schema.Boolean),
@@ -183,7 +172,6 @@ export type ToolCall = Schema.Schema.Type<typeof ToolCall>
 export const ToolResult = Schema.Struct({
  type: Schema.tag("tool-result"),
  id: ToolCallID,
-  itemId: Schema.optional(ResponseItemID),
  name: Schema.String,
  result: ToolResultValue,
  output: Schema.optional(ToolOutput),
@@ -195,7 +183,6 @@ export type ToolResult = Schema.Schema.Type<typeof ToolResult>
 export const ToolError = Schema.Struct({
  type: Schema.tag("tool-error"),
  id: ToolCallID,
-  itemId: Schema.optional(ResponseItemID),
  name: Schema.String,
  message: Schema.String,
  error: Schema.optional(Schema.Defect()),
@@ -347,14 +334,12 @@ const responseUsage = (events: ReadonlyArray<LLMEvent>) =>
 interface ContentAssembly {
  readonly contentIndex: number
  readonly text: string
-  readonly itemId?: ResponseItemID
  readonly providerMetadata?: ProviderMetadata
 }

 interface ToolInputAssembly {
  readonly name: string
  readonly text: string
-  readonly itemId?: ResponseItemID
  readonly providerMetadata?: ProviderMetadata
 }

@@ -400,27 +385,11 @@ const appendEvent = (state: ResponseState, event: LLMEvent): ResponseState => {
  }
 }

-const textContent = (
-  text: string,
-  itemId: ResponseItemID | undefined,
-  providerMetadata: ProviderMetadata | undefined,
-): ContentPart => ({
-  type: "text",
-  text,
-  ...(itemId === undefined ? {} : { itemId }),
-  ...(providerMetadata === undefined ? {} : { providerMetadata }),
-})
+const textContent = (text: string, providerMetadata: ProviderMetadata | undefined): ContentPart =>
+  providerMetadata === undefined ? { type: "text", text } : { type: "text", text, providerMetadata }

-const reasoningContent = (
-  text: string,
-  itemId: ResponseItemID | undefined,
-  providerMetadata: ProviderMetadata | undefined,
-): ContentPart => ({
-  type: "reasoning",
-  text,
-  ...(itemId === undefined ? {} : { itemId }),
-  ...(providerMetadata === undefined ? {} : { providerMetadata }),
-})
+const reasoningContent = (text: string, providerMetadata: ProviderMetadata | undefined): ContentPart =>
+  providerMetadata === undefined ? { type: "reasoning", text } : { type: "reasoning", text, providerMetadata }

 const contentWith = (state: ResponseState, content: ReadonlyArray<ContentPart>): ResponseState => ({
  ...state,
@@ -435,32 +404,26 @@ const replaceContent = (state: ResponseState, index: number, part: ContentPart)
    state.message.content.map((item, itemIndex) => (itemIndex === index ? part : item)),
  )

-const ensureText = (
-  state: ResponseState,
-  id: string,
-  itemId?: ResponseItemID,
-  providerMetadata?: ProviderMetadata,
-): ResponseState => {
+const ensureText = (state: ResponseState, id: string, providerMetadata?: ProviderMetadata): ResponseState => {
  if (state.textParts[id]) return state
  return {
-    ...appendContent(state, textContent("", itemId, providerMetadata)),
+    ...appendContent(state, textContent("", providerMetadata)),
    textParts: {
      ...state.textParts,
-      [id]: { contentIndex: state.message.content.length, text: "", itemId, providerMetadata },
+      [id]: { contentIndex: state.message.content.length, text: "", providerMetadata },
    },
  }
 }

 const reduceTextDelta = (state: ResponseState, event: TextDelta): ResponseState => {
-  const started = ensureText(state, event.id, event.itemId, event.providerMetadata)
+  const started = ensureText(state, event.id, event.providerMetadata)
  const current = started.textParts[event.id]
  if (!current) return started
  const text = current.text + event.text
-  const itemId = event.itemId ?? current.itemId
  const providerMetadata = event.providerMetadata ?? current.providerMetadata
  return {
-    ...replaceContent(started, current.contentIndex, textContent(text, itemId, providerMetadata)),
-    textParts: { ...started.textParts, [event.id]: { ...current, text, itemId, providerMetadata } },
+    ...replaceContent(started, current.contentIndex, textContent(text, providerMetadata)),
+    textParts: { ...started.textParts, [event.id]: { ...current, text, providerMetadata } },
  }
 }

@@ -468,39 +431,32 @@ const reduceTextEnd = (state: ResponseState, event: TextEnd): ResponseState => {
  const current = state.textParts[event.id]
  if (!current) return state
  const providerMetadata = event.providerMetadata ?? current.providerMetadata
-  const itemId = event.itemId ?? current.itemId
  return {
-    ...replaceContent(state, current.contentIndex, textContent(current.text, itemId, providerMetadata)),
-    textParts: { ...state.textParts, [event.id]: { ...current, itemId, providerMetadata } },
+    ...replaceContent(state, current.contentIndex, textContent(current.text, providerMetadata)),
+    textParts: { ...state.textParts, [event.id]: { ...current, providerMetadata } },
  }
 }

-const ensureReasoning = (
-  state: ResponseState,
-  id: string,
-  itemId?: ResponseItemID,
-  providerMetadata?: ProviderMetadata,
-): ResponseState => {
+const ensureReasoning = (state: ResponseState, id: string, providerMetadata?: ProviderMetadata): ResponseState => {
  if (state.reasoningParts[id]) return state
  return {
-    ...appendContent(state, reasoningContent("", itemId, providerMetadata)),
+    ...appendContent(state, reasoningContent("", providerMetadata)),
    reasoningParts: {
      ...state.reasoningParts,
-      [id]: { contentIndex: state.message.content.length, text: "", itemId, providerMetadata },
+      [id]: { contentIndex: state.message.content.length, text: "", providerMetadata },
    },
  }
 }

 const reduceReasoningDelta = (state: ResponseState, event: ReasoningDelta): ResponseState => {
-  const started = ensureReasoning(state, event.id, event.itemId, event.providerMetadata)
+  const started = ensureReasoning(state, event.id, event.providerMetadata)
  const current = started.reasoningParts[event.id]
  if (!current) return started
  const text = current.text + event.text
-  const itemId = event.itemId ?? current.itemId
  const providerMetadata = event.providerMetadata ?? current.providerMetadata
  return {
-    ...replaceContent(started, current.contentIndex, reasoningContent(text, itemId, providerMetadata)),
-    reasoningParts: { ...started.reasoningParts, [event.id]: { ...current, text, itemId, providerMetadata } },
+    ...replaceContent(started, current.contentIndex, reasoningContent(text, providerMetadata)),
+    reasoningParts: { ...started.reasoningParts, [event.id]: { ...current, text, providerMetadata } },
  }
 }

@@ -508,10 +464,9 @@ const reduceReasoningEnd = (state: ResponseState, event: ReasoningEnd): Response
  const current = state.reasoningParts[event.id]
  if (!current) return state
  const providerMetadata = event.providerMetadata ?? current.providerMetadata
-  const itemId = event.itemId ?? current.itemId
  return {
-    ...replaceContent(state, current.contentIndex, reasoningContent(current.text, itemId, providerMetadata)),
-    reasoningParts: { ...state.reasoningParts, [event.id]: { ...current, itemId, providerMetadata } },
+    ...replaceContent(state, current.contentIndex, reasoningContent(current.text, providerMetadata)),
+    reasoningParts: { ...state.reasoningParts, [event.id]: { ...current, providerMetadata } },
  }
 }

@@ -519,7 +474,7 @@ const reduceToolInputStart = (state: ResponseState, event: ToolInputStart): Resp
  ...state,
  toolInputs: {
    ...state.toolInputs,
-    [event.id]: { name: event.name, text: "", itemId: event.itemId, providerMetadata: event.providerMetadata },
+    [event.id]: { name: event.name, text: "", providerMetadata: event.providerMetadata },
  },
 })

@@ -540,7 +495,6 @@ const reduceToolInputEnd = (state: ResponseState, event: ToolInputEnd): Response
      [event.id]: {
        ...current,
        name: event.name,
-        itemId: event.itemId ?? current.itemId,
        providerMetadata: event.providerMetadata ?? current.providerMetadata,
      },
    },
@@ -550,7 +504,6 @@ const reduceToolInputEnd = (state: ResponseState, event: ToolInputEnd): Response
 const toolCallContent = (event: ToolCall): ContentPart =>
  ToolCallPart.make({
    id: event.id,
-    ...(event.itemId === undefined ? {} : { itemId: event.itemId }),
    name: event.name,
    input: event.input,
    ...(event.providerExecuted === undefined ? {} : { providerExecuted: event.providerExecuted }),
@@ -560,7 +513,6 @@ const toolCallContent = (event: ToolCall): ContentPart =>
 const toolResultContent = (event: ToolResult): ContentPart =>
  ToolResultPart.make({
    id: event.id,
-    ...(event.itemId === undefined ? {} : { itemId: event.itemId }),
    name: event.name,
    result: event.result,
    ...(event.providerExecuted === undefined ? {} : { providerExecuted: event.providerExecuted }),
@@ -576,13 +528,13 @@ const reduceResponseState = (state: ResponseState, event: LLMEvent): ResponseSta
  const next = appendEvent(state, event)
  switch (event.type) {
    case "text-start":
-      return ensureText(next, event.id, event.itemId, event.providerMetadata)
+      return ensureText(next, event.id, event.providerMetadata)
    case "text-delta":
      return reduceTextDelta(next, event)
    case "text-end":
      return reduceTextEnd(next, event)
    case "reasoning-start":
-      return ensureReasoning(next, event.id, event.itemId, event.providerMetadata)
+      return ensureReasoning(next, event.id, event.providerMetadata)
    case "reasoning-delta":
      return reduceReasoningDelta(next, event)
    case "reasoning-end":
@@ -21,9 +21,6 @@ export type ProviderID = typeof ProviderID.Type
 export const ResponseID = Schema.String
 export type ResponseID = Schema.Schema.Type<typeof ResponseID>

-export const ResponseItemID = Schema.String
-export type ResponseItemID = Schema.Schema.Type<typeof ResponseItemID>
-
 export const ContentBlockID = Schema.String
 export type ContentBlockID = Schema.Schema.Type<typeof ContentBlockID>

@@ -1,6 +1,6 @@
 import { Schema } from "effect"
 import { Tool } from "@opencode-ai/schema/tool"
-import { JsonSchema, MessageRole, ProviderMetadata, ResponseItemID } from "./ids"
+import { JsonSchema, MessageRole, ProviderMetadata } from "./ids"
 import { CacheHint, CachePolicy, GenerationOptions, HttpOptions, LanguageModelSchema, ProviderOptions } from "./options"
 import { isRecord } from "../utils/record"

@@ -25,7 +25,6 @@ export const SystemPart = Object.assign(systemPartSchema, {
 export const TextPart = Schema.Struct({
  type: Schema.Literal("text"),
  text: Schema.String,
-  itemId: Schema.optional(ResponseItemID),
  cache: Schema.optional(CacheHint),
  metadata: Schema.optional(Schema.Record(Schema.String, Schema.Unknown)),
  providerMetadata: Schema.optional(ProviderMetadata),
@@ -122,7 +121,6 @@ export const ToolCallPart = Object.assign(
  Schema.Struct({
    type: Schema.Literal("tool-call"),
    id: Schema.String,
-    itemId: Schema.optional(ResponseItemID),
    name: Schema.String,
    input: Schema.Unknown,
    providerExecuted: Schema.optional(Schema.Boolean),
@@ -140,7 +138,6 @@ export const ToolResultPart = Object.assign(
  Schema.Struct({
    type: Schema.Literal("tool-result"),
    id: Schema.String,
-    itemId: Schema.optional(ResponseItemID),
    name: Schema.String,
    result: ToolResultValue,
    providerExecuted: Schema.optional(Schema.Boolean),
@@ -157,7 +154,6 @@ export const ToolResultPart = Object.assign(
    ): ToolResultPart => ({
      type: "tool-result",
      id: input.id,
-      ...(input.itemId === undefined ? {} : { itemId: input.itemId }),
      name: input.name,
      result: ToolResultValue.make(input.result, input.resultType),
      providerExecuted: input.providerExecuted,
@@ -172,7 +168,6 @@ export type ToolResultPart = Schema.Schema.Type<typeof ToolResultPart>
 export const ReasoningPart = Schema.Struct({
  type: Schema.Literal("reasoning"),
  text: Schema.String,
-  itemId: Schema.optional(ResponseItemID),
  encrypted: Schema.optional(Schema.String),
  cache: Schema.optional(CacheHint),
  metadata: Schema.optional(Schema.Record(Schema.String, Schema.Unknown)),
@@ -186,7 +181,7 @@ export const ContentPart = Schema.Union([TextPart, MediaPart, ToolCallPart, Tool
 export type ContentPart = Schema.Schema.Type<typeof ContentPart>

 export class Message extends Schema.Class<Message>("LLM.Message")({
-  id: Schema.optional(ResponseItemID),
+  id: Schema.optional(Schema.String),
  role: MessageRole,
  content: Schema.Array(ContentPart),
  metadata: Schema.optional(Schema.Record(Schema.String, Schema.Unknown)),
@@ -79,7 +79,7 @@ const result = (call: ToolCallPart, value: ToolResultValueType | ToolSettlement,
              id: call.id,
              name: call.name,
              result: settlement.result,
-              ...(call.providerMetadata === undefined ? {} : { providerMetadata: call.providerMetadata }),
+              providerMetadata: call.providerMetadata,
            }),
          ]
        : [
@@ -88,7 +88,7 @@ const result = (call: ToolCallPart, value: ToolResultValueType | ToolSettlement,
              name: call.name,
              result: settlement.result,
              output: settlement.output,
-              ...(call.providerMetadata === undefined ? {} : { providerMetadata: call.providerMetadata }),
+              providerMetadata: call.providerMetadata,
            }),
          ],
  }
@@ -44,7 +44,7 @@
        "headers": {
          "content-type": "application/json"
        },
-        "body": "{\"model\":\"gpt-5.5\",\"input\":[{\"role\":\"user\",\"content\":[{\"type\":\"input_text\",\"text\":\"Think briefly, then reply exactly with: Hello!\"}]},{\"type\":\"reasoning\",\"summary\":[],\"encrypted_content\":\"gAAAAABqEjXoGMCw3WDXpoD9151PEr2Lt8raW7KBKefQhZJGWx5f8jy152bApO6oE-Mr1BhUtfZNq3OPBVfSL4ioQ9bHREfujIBXgk9LUDBAz2Sle7KjOr9HaUV16A4HBiaFIRFjsHPS9G8yEySp1m6F1CD_WR6apyUGgugRh_y39EcOJmxPOzmiac5DVM6fraA1VpcGbqrZ1x2ANHFDOfnYTycPtPNTgzE7LjkYjDDWbT03uN1YxfP4pqjDVRzY14pA8bSZ8ys-pDv5kUFCAsw-OlU4jYKUXp-M8_6KTaRQP71LPwppt__zG_NJPfy-qUil4pOU8_NoxtxerHgLLXbfExZdzfpoGinoEjn7nj7BJDEtl-LNeNEb5c-1ZymNfVMp-Cs3fLEPkAV8rtHFtZ0MhE_07GKbGo7hTrOmkM4DydxmHsdWGNbXAG35cprslEA5P7p3GHFKnRs5hGs2eq-XcZ3yki64ZBOU_Tv6UR7nUH09gF1rdrJo3dpre6M00COwwdZ02zUP5KxCuI8FKu2jsZu9zgMVXDALsdtM5orTCVLXsn4rddWd111zE-vMjNmMMmktW2cHMjH7j1ooA-9P083koNVYiLi4UhMA64gTqgyl8MxkZekl7eFSMa7qk295NaHOKtFxzYYcZ9jdioCwSPSZ0ZZWLoNgrK7SWfRh0uaTHNcMZ3wq8ae6CguktIeVTCPTQAqJLQqd7AU0oOCKCJ7BWnC-L8UC6m7Pm9ZS958uUVeWBhgKHzMAGq9UeQB7IEeAcbMn3EDgOSfd8qCb8iwU9iG9dcu9axQwWU7pd7kd-T-He61W7z5wWgpx1KehWCxrN6kuKSo6p-uUfwVnJukreOn8BJNAzADQgz68bhmN9VGih7YcKVnLgwDwKditrjSd6-tfE0Baarj3jWENvT6ohY17R9FDrKS-2v8IIX6tGjoKJw8SRhaWLNv4vWlmxRgR0gdac3qumd0GKqsWSveNz01naA==\",\"id\":\"rs_0a0794dab3b8ec7d016a1235e7ce3881958a5eca32a36a14c5\"},{\"role\":\"assistant\",\"content\":[{\"type\":\"output_text\",\"text\":\"Hello!\"}]},{\"role\":\"user\",\"content\":[{\"type\":\"input_text\",\"text\":\"Now reply exactly with: Done.\"}]}],\"store\":false,\"include\":[\"reasoning.encrypted_content\"],\"reasoning\":{\"effort\":\"low\",\"summary\":\"auto\"},\"text\":{\"verbosity\":\"low\"},\"max_output_tokens\":40,\"stream\":true}"
+        "body": "{\"model\":\"gpt-5.5\",\"input\":[{\"role\":\"user\",\"content\":[{\"type\":\"input_text\",\"text\":\"Think briefly, then reply exactly with: Hello!\"}]},{\"type\":\"reasoning\",\"summary\":[],\"encrypted_content\":\"gAAAAABqEjXoGMCw3WDXpoD9151PEr2Lt8raW7KBKefQhZJGWx5f8jy152bApO6oE-Mr1BhUtfZNq3OPBVfSL4ioQ9bHREfujIBXgk9LUDBAz2Sle7KjOr9HaUV16A4HBiaFIRFjsHPS9G8yEySp1m6F1CD_WR6apyUGgugRh_y39EcOJmxPOzmiac5DVM6fraA1VpcGbqrZ1x2ANHFDOfnYTycPtPNTgzE7LjkYjDDWbT03uN1YxfP4pqjDVRzY14pA8bSZ8ys-pDv5kUFCAsw-OlU4jYKUXp-M8_6KTaRQP71LPwppt__zG_NJPfy-qUil4pOU8_NoxtxerHgLLXbfExZdzfpoGinoEjn7nj7BJDEtl-LNeNEb5c-1ZymNfVMp-Cs3fLEPkAV8rtHFtZ0MhE_07GKbGo7hTrOmkM4DydxmHsdWGNbXAG35cprslEA5P7p3GHFKnRs5hGs2eq-XcZ3yki64ZBOU_Tv6UR7nUH09gF1rdrJo3dpre6M00COwwdZ02zUP5KxCuI8FKu2jsZu9zgMVXDALsdtM5orTCVLXsn4rddWd111zE-vMjNmMMmktW2cHMjH7j1ooA-9P083koNVYiLi4UhMA64gTqgyl8MxkZekl7eFSMa7qk295NaHOKtFxzYYcZ9jdioCwSPSZ0ZZWLoNgrK7SWfRh0uaTHNcMZ3wq8ae6CguktIeVTCPTQAqJLQqd7AU0oOCKCJ7BWnC-L8UC6m7Pm9ZS958uUVeWBhgKHzMAGq9UeQB7IEeAcbMn3EDgOSfd8qCb8iwU9iG9dcu9axQwWU7pd7kd-T-He61W7z5wWgpx1KehWCxrN6kuKSo6p-uUfwVnJukreOn8BJNAzADQgz68bhmN9VGih7YcKVnLgwDwKditrjSd6-tfE0Baarj3jWENvT6ohY17R9FDrKS-2v8IIX6tGjoKJw8SRhaWLNv4vWlmxRgR0gdac3qumd0GKqsWSveNz01naA==\"},{\"role\":\"assistant\",\"content\":[{\"type\":\"output_text\",\"text\":\"Hello!\"}]},{\"role\":\"user\",\"content\":[{\"type\":\"input_text\",\"text\":\"Now reply exactly with: Done.\"}]}],\"store\":false,\"include\":[\"reasoning.encrypted_content\"],\"reasoning\":{\"effort\":\"low\",\"summary\":\"auto\"},\"text\":{\"verbosity\":\"low\"},\"max_output_tokens\":40,\"stream\":true}"
      },
      "response": {
        "status": 200,
@@ -8,6 +8,7 @@ import {
  type ProviderMetadata,
  type ToolCallPart,
  ToolResultPart,
+  type ToolResultValue,
  type Usage,
 } from "../../src/schema"
 import { type Tools, toDefinitions } from "../../src/tool"
@@ -60,10 +61,9 @@ export const runTools = <T extends Tools>(options: RunOptions<T>) =>
            ...dispatched.map(([call, dispatched]) =>
              Message.tool({
                id: call.id,
-                itemId: dispatched.events.find(LLMEvent.is.toolResult)?.itemId,
                name: call.name,
                result: dispatched.result,
-                providerMetadata: dispatched.events.find(LLMEvent.is.toolResult)?.providerMetadata,
+                providerMetadata: call.providerMetadata,
              }),
            ),
          ],
@@ -89,15 +89,9 @@ const stepState = (events: ReadonlyArray<LLMEvent>) => {

  for (const event of events) {
    if (event.type === "text-delta" || event.type === "reasoning-delta") {
-      appendText(assistantContent, event.type === "text-delta" ? "text" : "reasoning", event.text, event.itemId)
+      appendText(assistantContent, event.type === "text-delta" ? "text" : "reasoning", event.text)
    } else if (event.type === "text-end" || event.type === "reasoning-end") {
-      appendText(
-        assistantContent,
-        event.type === "text-end" ? "text" : "reasoning",
-        "",
-        event.itemId,
-        event.providerMetadata,
-      )
+      appendText(assistantContent, event.type === "text-end" ? "text" : "reasoning", "", event.providerMetadata)
    } else if (event.type === "tool-call") {
      assistantContent.push(event)
      if (!event.providerExecuted) toolCalls.push(event)
@@ -105,7 +99,6 @@ const stepState = (events: ReadonlyArray<LLMEvent>) => {
      assistantContent.push(
        ToolResultPart.make({
          id: event.id,
-          itemId: event.itemId,
          name: event.name,
          result: event.result,
          providerExecuted: true,
@@ -125,7 +118,6 @@ const appendText = (
  content: ContentPart[],
  type: "text" | "reasoning",
  text: string,
-  itemId?: string,
  providerMetadata?: ProviderMetadata,
 ) => {
  const last = content.at(-1)
@@ -133,12 +125,11 @@ const appendText = (
    content[content.length - 1] = {
      ...last,
      text: `${last.text}${text}`,
-      itemId: itemId ?? last.itemId,
      providerMetadata: providerMetadata ?? last.providerMetadata,
    }
    return
  }
-  content.push({ type, text, itemId, providerMetadata })
+  content.push({ type, text, providerMetadata })
 }

 const addUsage = (left: Usage | undefined, right: Usage | undefined): Usage | undefined => {
@@ -16,13 +16,6 @@ const model = Gemini.route
  })
  .model({ id: "gemini-2.5-flash" })

-const gemini3 = Gemini.route
-  .with({
-    endpoint: { baseURL: "https://generativelanguage.test/v1beta/" },
-    auth: Auth.header("x-goog-api-key", "test"),
-  })
-  .model({ id: "gemini-3-flash-preview" })
-
 const request = LLM.request({
  id: "req_1",
  model,
@@ -93,39 +86,6 @@ describe("Gemini route", () => {
    }),
  )

-  it.effect("forwards standard Gemini generation options", () =>
-    Effect.gen(function* () {
-      const prepared = yield* compileRequest(
-        LLM.request({
-          model,
-          prompt: "Say hello.",
-          generation: {
-            maxTokens: 40,
-            temperature: 0.2,
-            topP: 0.8,
-            topK: 12,
-            frequencyPenalty: 0.3,
-            presencePenalty: 0.4,
-            seed: 42,
-            stop: ["done"],
-          },
-        }),
-      )
-
-      expect(prepared.body.generationConfig).toEqual({
-        maxOutputTokens: 40,
-        temperature: 0.2,
-        topP: 0.8,
-        topK: 12,
-        frequencyPenalty: 0.3,
-        presencePenalty: 0.4,
-        seed: 42,
-        stopSequences: ["done"],
-        thinkingConfig: undefined,
-      })
-    }),
-  )
-
  it.effect("lowers chronological system updates to wrapped user text in order", () =>
    Effect.gen(function* () {
      const prepared = yield* compileRequest(
@@ -390,100 +350,6 @@ describe("Gemini route", () => {
    }),
  )

-  it.effect("preserves nested empty object tool schemas", () =>
-    Effect.gen(function* () {
-      const prepared = yield* compileRequest(
-        LLM.request({
-          model,
-          prompt: "Use the tool.",
-          tools: [
-            {
-              name: "configure",
-              description: "Configure the operation",
-              inputSchema: {
-                type: "object",
-                required: ["options"],
-                properties: {
-                  options: { type: "object", description: "Optional provider settings", properties: {} },
-                },
-              },
-            },
-          ],
-        }),
-      )
-
-      expect(prepared.body.tools).toEqual([
-        {
-          functionDeclarations: [
-            {
-              name: "configure",
-              description: "Configure the operation",
-              parameters: {
-                type: "object",
-                required: ["options"],
-                properties: {
-                  options: { type: "object", description: "Optional provider settings", properties: {} },
-                },
-              },
-            },
-          ],
-        },
-      ])
-    }),
-  )
-
-  it.effect("projects Gemini type arrays without narrowing their allowed values", () =>
-    Effect.gen(function* () {
-      const prepared = yield* compileRequest(
-        LLM.request({
-          model,
-          prompt: "Use the tool.",
-          tools: [
-            {
-              name: "filter",
-              description: "Filter values",
-              inputSchema: {
-                type: "object",
-                properties: {
-                  status: { type: ["number", "string"], description: "Status filter" },
-                  maybe: { type: ["string", "null"] },
-                  nothing: { type: ["null"] },
-                  explicit: { anyOf: [{ type: "string" }, { type: "null" }] },
-                  choice: { anyOf: [{ type: "string" }, { type: "number" }, { type: "null" }] },
-                },
-              },
-            },
-          ],
-        }),
-      )
-
-      expect(prepared.body.tools?.[0]?.functionDeclarations[0]?.parameters).toEqual({
-        type: "object",
-        properties: {
-          status: {
-            description: "Status filter",
-            anyOf: [{ type: "number" }, { type: "string" }],
-          },
-          maybe: {
-            nullable: true,
-            anyOf: [{ type: "string" }],
-          },
-          nothing: {
-            type: "null",
-          },
-          explicit: {
-            type: "string",
-            nullable: true,
-          },
-          choice: {
-            anyOf: [{ type: "string" }, { type: "number" }],
-            nullable: true,
-          },
-        },
-      })
-    }),
-  )
-
  it.effect("parses text, reasoning, and usage stream fixtures", () =>
    Effect.gen(function* () {
      const body = sseEvents(
@@ -670,44 +536,6 @@ describe("Gemini route", () => {
    }),
  )

-  it.effect("replays unsigned Gemini 3 tool calls with the validator bypass sentinel", () =>
-    Effect.gen(function* () {
-      const prepared = yield* compileRequest(
-        LLM.request({
-          model: gemini3,
-          messages: [
-            Message.assistant([ToolCallPart.make({ id: "tool_0", name: "lookup", input: { query: "weather" } })]),
-            Message.tool({ id: "tool_0", name: "lookup", result: "done", resultType: "text" }),
-          ],
-        }),
-      )
-
-      expect(prepared.body.contents).toEqual([
-        {
-          role: "model",
-          parts: [
-            {
-              functionCall: { id: undefined, name: "lookup", args: { query: "weather" } },
-              thoughtSignature: "skip_thought_signature_validator",
-            },
-          ],
-        },
-        {
-          role: "user",
-          parts: [
-            {
-              functionResponse: {
-                id: undefined,
-                name: "lookup",
-                response: { name: "lookup", content: "done" },
-              },
-            },
-          ],
-        },
-      ])
-    }),
-  )
-
  it.effect("emits streamed tool calls and maps finish reason", () =>
    Effect.gen(function* () {
      const body = sseEvents({
@@ -45,7 +45,7 @@ describe("Open Responses-compatible route", () => {
          },
        },
      })
-      expect(prepared.body).toMatchObject({
+      expect(prepared.body).toEqual({
        model: "example-model",
        input: [
          { role: "system", content: "You are concise." },
@@ -53,8 +53,6 @@ describe("Open Responses-compatible route", () => {
        ],
        stream: true,
      })
-      expect(prepared.body.input[0]).not.toHaveProperty("id")
-      expect(prepared.body.input[1]).not.toHaveProperty("id")
    }),
  )

@@ -69,9 +69,6 @@ describe("OpenAI Responses route", () => {
        stream: true,
        max_output_tokens: 20,
        temperature: 0,
-        tool_choice: undefined,
-        tools: undefined,
-        top_p: undefined,
      })
    }),
  )
@@ -332,7 +329,7 @@ describe("OpenAI Responses route", () => {
      yield* LLMClient.generate(
        LLMRequest.update(request, {
          model: Azure.configure({
-            baseURL: "https://opencode-test.openai.azure.com/openai/",
+            baseURL: "https://opencode-test.openai.azure.com/openai/v1/",
            apiKey: "azure-key",
            headers: { authorization: "Bearer stale" },
          }).responses("gpt-4.1-mini"),
@@ -413,7 +410,7 @@ describe("OpenAI Responses route", () => {
        }),
      )

-      expect(prepared.body).toMatchObject({
+      expect(prepared.body).toEqual({
        model: "gpt-4.1-mini",
        input: [
          { role: "user", content: [{ type: "input_text", text: "What is the weather?" }] },
@@ -428,60 +425,6 @@ describe("OpenAI Responses route", () => {
        tools: undefined,
        top_p: undefined,
      })
-      const call = prepared.body.input.find((item) => "type" in item && item.type === "function_call")
-      const output = prepared.body.input.find((item) => "type" in item && item.type === "function_call_output")
-      expect(call?.id).toBeUndefined()
-      expect(output?.id).toBeUndefined()
-    }),
-  )
-
-  it.effect("does not generate response item ids for client-created history", () =>
-    Effect.sync(() => {
-      const canonical = LLM.request({
-        model,
-        messages: [
-          Message.assistant([
-            { type: "text", text: "Working." },
-            { type: "reasoning", text: "Thinking." },
-            ToolCallPart.make({ id: "call_1", name: "lookup", input: {} }),
-          ]),
-          Message.tool({ id: "call_1", name: "lookup", result: "done" }),
-        ],
-      })
-
-      expect(canonical.messages[0]?.id).toBeUndefined()
-      expect(canonical.messages[1]?.id).toBeUndefined()
-      expect(canonical.messages[0]?.content.every((part) => part.type === "media" || part.itemId === undefined)).toBe(
-        true,
-      )
-      expect(canonical.messages[1]?.content[0]).not.toHaveProperty("itemId")
-    }),
-  )
-
-  it.effect("preserves opaque assistant item ids without assigning ids to function items", () =>
-    Effect.gen(function* () {
-      const canonical = LLM.request({
-        model,
-        messages: [
-          Message.assistant([
-            { type: "text", text: "Calling.", itemId: "plain-text" },
-            ToolCallPart.make({ id: "call_1", itemId: "plain-call", name: "lookup", input: {} }),
-          ]),
-          Message.tool({ id: "call_1", itemId: "plain-output", name: "lookup", result: "done" }),
-        ],
-      })
-      const prepared = yield* compileRequest(canonical)
-
-      expect(canonical.messages[0]?.content.map((part) => (part.type === "media" ? undefined : part.itemId))).toEqual([
-        "plain-text",
-        "plain-call",
-      ])
-      expect(canonical.messages[1]?.content[0]).toMatchObject({ itemId: "plain-output" })
-      expect(prepared.body.input).toEqual([
-        { role: "assistant", id: "plain-text", content: [{ type: "output_text", text: "Calling." }] },
-        { type: "function_call", call_id: "call_1", name: "lookup", arguments: "{}" },
-        { type: "function_call_output", call_id: "call_1", output: '"done"' },
-      ])
    }),
  )

@@ -921,21 +864,9 @@ describe("OpenAI Responses route", () => {
      expect(response.text).toBe("Hello!")
      expect(response.events).toEqual([
        { type: "step-start", index: 0 },
-        { type: "text-start", id: "msg_1", itemId: "msg_1", providerMetadata: { openai: { itemId: "msg_1" } } },
-        {
-          type: "text-delta",
-          id: "msg_1",
-          itemId: "msg_1",
-          text: "Hello",
-          providerMetadata: { openai: { itemId: "msg_1" } },
-        },
-        {
-          type: "text-delta",
-          id: "msg_1",
-          itemId: "msg_1",
-          text: "!",
-          providerMetadata: { openai: { itemId: "msg_1" } },
-        },
+        { type: "text-start", id: "msg_1" },
+        { type: "text-delta", id: "msg_1", text: "Hello" },
+        { type: "text-delta", id: "msg_1", text: "!" },
        { type: "text-end", id: "msg_1" },
        {
          type: "step-finish",
@@ -992,20 +923,17 @@ describe("OpenAI Responses route", () => {
        {
          type: "text",
          text: "Checking.",
-          itemId: "msg_commentary",
-          providerMetadata: { openai: { itemId: "msg_commentary", phase: "commentary" } },
+          providerMetadata: { openai: { phase: "commentary" } },
        },
        {
          type: "text",
          text: "Finished.",
-          itemId: "msg_final",
-          providerMetadata: { openai: { itemId: "msg_final", phase: "final_answer" } },
+          providerMetadata: { openai: { phase: "final_answer" } },
        },
        {
          type: "text",
          text: "Unclassified.",
-          itemId: "msg_null",
-          providerMetadata: { openai: { itemId: "msg_null", phase: null } },
+          providerMetadata: { openai: { phase: null } },
        },
      ])

@@ -1013,19 +941,16 @@ describe("OpenAI Responses route", () => {
      expect(prepared.body.input).toEqual([
        {
          role: "assistant",
-          id: "msg_commentary",
          content: [{ type: "output_text", text: "Checking." }],
          phase: "commentary",
        },
        {
          role: "assistant",
-          id: "msg_final",
          content: [{ type: "output_text", text: "Finished." }],
          phase: "final_answer",
        },
        {
          role: "assistant",
-          id: "msg_null",
          content: [{ type: "output_text", text: "Unclassified." }],
          phase: null,
        },
@@ -1118,24 +1043,12 @@ describe("OpenAI Responses route", () => {
      )

      expect(response.events.filter((event) => event.type.startsWith("text-"))).toEqual([
-        { type: "text-start", id: "msg_1", itemId: "msg_1", providerMetadata: { openai: { itemId: "msg_1" } } },
-        {
-          type: "text-delta",
-          id: "msg_1",
-          itemId: "msg_1",
-          text: "First",
-          providerMetadata: { openai: { itemId: "msg_1" } },
-        },
-        { type: "text-end", id: "msg_1", itemId: "msg_1", providerMetadata: { openai: { itemId: "msg_1" } } },
-        { type: "text-start", id: "msg_2", itemId: "msg_2", providerMetadata: { openai: { itemId: "msg_2" } } },
-        {
-          type: "text-delta",
-          id: "msg_2",
-          itemId: "msg_2",
-          text: "Second",
-          providerMetadata: { openai: { itemId: "msg_2" } },
-        },
-        { type: "text-end", id: "msg_2", itemId: "msg_2", providerMetadata: { openai: { itemId: "msg_2" } } },
+        { type: "text-start", id: "msg_1" },
+        { type: "text-delta", id: "msg_1", text: "First" },
+        { type: "text-end", id: "msg_1" },
+        { type: "text-start", id: "msg_2" },
+        { type: "text-delta", id: "msg_2", text: "Second" },
+        { type: "text-end", id: "msg_2" },
      ])
    }),
  )
@@ -1155,15 +1068,9 @@ describe("OpenAI Responses route", () => {
      expect(response.text).toBe("Hello")
      expect(response.events).toMatchObject([
        { type: "step-start", index: 0 },
-        { type: "reasoning-start", id: "rs_1", itemId: "rs_1", providerMetadata: { openai: { itemId: "rs_1" } } },
-        {
-          type: "reasoning-delta",
-          id: "rs_1",
-          itemId: "rs_1",
-          text: "thinking",
-          providerMetadata: { openai: { itemId: "rs_1" } },
-        },
-        { type: "text-start", id: "msg_1", itemId: "msg_1", providerMetadata: { openai: { itemId: "msg_1" } } },
+        { type: "reasoning-start", id: "rs_1" },
+        { type: "reasoning-delta", id: "rs_1", text: "thinking" },
+        { type: "text-start", id: "msg_1" },
        { type: "text-delta", id: "msg_1", text: "Hello" },
        { type: "reasoning-end", id: "rs_1" },
        { type: "text-end", id: "msg_1" },
@@ -1172,8 +1079,8 @@ describe("OpenAI Responses route", () => {
      ])
      expect(response.events.filter((event) => event.type === "finish")).toHaveLength(1)
      expect(response.message.content).toEqual([
-        { type: "reasoning", text: "thinking", itemId: "rs_1", providerMetadata: { openai: { itemId: "rs_1" } } },
-        { type: "text", text: "Hello", itemId: "msg_1", providerMetadata: { openai: { itemId: "msg_1" } } },
+        { type: "reasoning", text: "thinking" },
+        { type: "text", text: "Hello" },
      ])
    }),
  )
@@ -1204,7 +1111,6 @@ describe("OpenAI Responses route", () => {
        expect.objectContaining({
          type: "reasoning-end",
          id: "rs_1",
-          itemId: "rs_1",
          providerMetadata: { openai: { itemId: "rs_1", reasoningEncryptedContent: "encrypted-state" } },
        }),
      )
@@ -1245,34 +1151,19 @@ describe("OpenAI Responses route", () => {
        {
          type: "reasoning-start",
          id: "rs_1:0",
-          itemId: "rs_1",
          providerMetadata: { openai: { itemId: "rs_1", reasoningEncryptedContent: null } },
        },
-        {
-          type: "reasoning-delta",
-          id: "rs_1:0",
-          itemId: "rs_1",
-          text: "First",
-          providerMetadata: { openai: { itemId: "rs_1" } },
-        },
-        { type: "reasoning-end", id: "rs_1:0", itemId: "rs_1", providerMetadata: { openai: { itemId: "rs_1" } } },
+        { type: "reasoning-delta", id: "rs_1:0", text: "First" },
+        { type: "reasoning-end", id: "rs_1:0", providerMetadata: { openai: { itemId: "rs_1" } } },
        {
          type: "reasoning-start",
          id: "rs_1:1",
-          itemId: "rs_1",
          providerMetadata: { openai: { itemId: "rs_1", reasoningEncryptedContent: null } },
        },
-        {
-          type: "reasoning-delta",
-          id: "rs_1:1",
-          itemId: "rs_1",
-          text: "Second",
-          providerMetadata: { openai: { itemId: "rs_1" } },
-        },
+        { type: "reasoning-delta", id: "rs_1:1", text: "Second" },
        {
          type: "reasoning-end",
          id: "rs_1:1",
-          itemId: "rs_1",
          providerMetadata: { openai: { itemId: "rs_1", reasoningEncryptedContent: "encrypted-state" } },
        },
        { type: "step-finish", index: 0, reason: { normalized: "stop", raw: undefined } },
@@ -1310,8 +1201,8 @@ describe("OpenAI Responses route", () => {
      )

      expect(response.events.filter((event) => event.type === "reasoning-end")).toEqual([
-        { type: "reasoning-end", id: "rs_1:0", itemId: "rs_1", providerMetadata: { openai: { itemId: "rs_1" } } },
-        { type: "reasoning-end", id: "rs_1:1", itemId: "rs_1", providerMetadata: { openai: { itemId: "rs_1" } } },
+        { type: "reasoning-end", id: "rs_1:0", providerMetadata: { openai: { itemId: "rs_1" } } },
+        { type: "reasoning-end", id: "rs_1:1", providerMetadata: { openai: { itemId: "rs_1" } } },
      ])
    }),
  )
@@ -1359,7 +1250,7 @@ describe("OpenAI Responses route", () => {
                  { role: "user", content: [{ type: "input_text", text: "Summarize it." }] },
                ],
              })
-              expect(body.input[1]).toHaveProperty("id", "rs_1")
+              expect(body.input[1]).not.toHaveProperty("id")
              return input.respond(
                sseEvents(
                  { type: "response.output_text.delta", item_id: "msg_1", delta: "Parser now round-trips reasoning." },
@@ -1406,7 +1297,6 @@ describe("OpenAI Responses route", () => {
        { role: "assistant", content: [{ type: "output_text", text: "Before." }] },
        {
          type: "reasoning",
-          id: "rs_1",
          encrypted_content: "encrypted-state",
          summary: [{ type: "summary_text", text: "Checked order." }],
        },
@@ -1415,7 +1305,7 @@ describe("OpenAI Responses route", () => {
    }),
  )

-  it.effect("replays complete stored reasoning items with their id", () =>
+  it.effect("references stored reasoning items by id", () =>
    Effect.gen(function* () {
      const prepared = yield* compileRequest(
        LLM.request({
@@ -1433,14 +1323,7 @@ describe("OpenAI Responses route", () => {
        }),
      )

-      expect(prepared.body.input).toEqual([
-        {
-          type: "reasoning",
-          id: "rs_1",
-          summary: [{ type: "summary_text", text: "Checked the previous diff." }],
-          encrypted_content: undefined,
-        },
-      ])
+      expect(prepared.body.input).toEqual([{ type: "item_reference", id: "rs_1" }])
    }),
  )

@@ -1549,7 +1432,6 @@ describe("OpenAI Responses route", () => {
      expect(prepared.body.input).toEqual([
        {
          type: "reasoning",
-          id: "rs_1",
          encrypted_content: "encrypted-state",
          summary: [
            { type: "summary_text", text: "First" },
@@ -1629,7 +1511,6 @@ describe("OpenAI Responses route", () => {
        outputTokens: 1,
        nonCachedInputTokens: 5,
        cacheReadInputTokens: undefined,
-        cacheWriteInputTokens: undefined,
        reasoningTokens: undefined,
        totalTokens: 6,
        providerMetadata: { openai: { input_tokens: 5, output_tokens: 1 } },
@@ -1640,35 +1521,30 @@ describe("OpenAI Responses route", () => {
        {
          type: "tool-input-start",
          id: "call_1",
-          itemId: "item_1",
          name: "lookup",
          providerMetadata: { openai: { itemId: "item_1" } },
        },
        {
          type: "tool-input-delta",
          id: "call_1",
-          itemId: "item_1",
          name: "lookup",
          text: '{"query"',
        },
        {
          type: "tool-input-delta",
          id: "call_1",
-          itemId: "item_1",
          name: "lookup",
          text: ':"weather"}',
        },
        {
          type: "tool-input-end",
          id: "call_1",
-          itemId: "item_1",
          name: "lookup",
          providerMetadata: { openai: { itemId: "item_1" } },
        },
        {
          type: "tool-call",
          id: "call_1",
-          itemId: "item_1",
          name: "lookup",
          input: { query: "weather" },
          providerExecuted: undefined,
@@ -1688,17 +1564,6 @@ describe("OpenAI Responses route", () => {
          usage,
        },
      ])
-      expect(response.message.content).toEqual([
-        {
-          type: "tool-call",
-          id: "call_1",
-          itemId: "item_1",
-          name: "lookup",
-          input: { query: "weather" },
-          providerExecuted: undefined,
-          providerMetadata: { openai: { itemId: "item_1" } },
-        },
-      ])
    }),
  )

@@ -1731,7 +1596,6 @@ describe("OpenAI Responses route", () => {
      expect(response.events.find(LLMEvent.is.toolInputError)).toEqual({
        type: "tool-input-error",
        id: "call_1",
-        itemId: "item_1",
        name: "lookup",
        raw: '{"query":"partial',
      })
@@ -1788,7 +1652,6 @@ describe("OpenAI Responses route", () => {
        {
          type: "tool-call",
          id: "ws_1",
-          itemId: "ws_1",
          name: "web_search",
          input: { type: "search", query: "effect 4" },
          providerExecuted: true,
@@ -1797,35 +1660,11 @@ describe("OpenAI Responses route", () => {
        {
          type: "tool-result",
          id: "ws_1",
-          itemId: "ws_1",
          name: "web_search",
          result: { type: "json", value: item },
          providerExecuted: true,
-          providerMetadata: { openai: { itemId: "ws_1", item } },
-          output: undefined,
-        },
-      ])
-      expect(response.message.content).toEqual([
-        {
-          type: "tool-call",
-          id: "ws_1",
-          itemId: "ws_1",
-          name: "web_search",
-          input: { type: "search", query: "effect 4" },
-          providerExecuted: true,
          providerMetadata: { openai: { itemId: "ws_1" } },
        },
-        {
-          type: "tool-result",
-          id: "ws_1",
-          itemId: "ws_1",
-          name: "web_search",
-          result: { type: "json", value: item },
-          providerExecuted: true,
-          providerMetadata: { openai: { itemId: "ws_1", item } },
-          metadata: undefined,
-          cache: undefined,
-        },
      ])
    }),
  )
@@ -1903,7 +1742,6 @@ describe("OpenAI Responses route", () => {
      expect(toolCall).toEqual({
        type: "tool-call",
        id: "ci_1",
-        itemId: "ci_1",
        name: "code_interpreter",
        input: { code: "print(1+1)", container_id: "cnt_xyz" },
        providerExecuted: true,
@@ -1913,12 +1751,10 @@ describe("OpenAI Responses route", () => {
      expect(toolResult).toEqual({
        type: "tool-result",
        id: "ci_1",
-        itemId: "ci_1",
        name: "code_interpreter",
        result: { type: "json", value: item },
        providerExecuted: true,
-        providerMetadata: { openai: { itemId: "ci_1", item } },
-        output: undefined,
+        providerMetadata: { openai: { itemId: "ci_1" } },
      })
    }),
  )
@@ -49,43 +49,6 @@ describe("LLMResponse reducer", () => {
    expect(state.message.content).toEqual([{ type: "text", text: "partial" }])
  })

-  test("assembles response item identity and provider metadata", () => {
-    const response = LLMResponse.fromEvents([
-      LLMEvent.textStart({ id: "text-block", itemId: "msg_existing" }),
-      LLMEvent.textDelta({
-        id: "text-block",
-        itemId: "msg_existing",
-        text: "Answer",
-        providerMetadata: { openai: { itemId: "msg_existing" } },
-      }),
-      LLMEvent.textEnd({ id: "text-block", itemId: "msg_existing" }),
-      LLMEvent.reasoningStart({ id: "reasoning-block", itemId: "rs_existing" }),
-      LLMEvent.reasoningDelta({
-        id: "reasoning-block",
-        itemId: "rs_existing",
-        text: "Thought",
-        providerMetadata: { openai: { itemId: "rs_existing" } },
-      }),
-      LLMEvent.reasoningEnd({ id: "reasoning-block", itemId: "rs_existing" }),
-      LLMEvent.finish({ reason: { normalized: "stop" } }),
-    ])
-
-    expect(response?.message.content).toEqual([
-      {
-        type: "text",
-        text: "Answer",
-        itemId: "msg_existing",
-        providerMetadata: { openai: { itemId: "msg_existing" } },
-      },
-      {
-        type: "reasoning",
-        text: "Thought",
-        itemId: "rs_existing",
-        providerMetadata: { openai: { itemId: "rs_existing" } },
-      },
-    ])
-  })
-
  test("does not complete ended content without a terminal finish", () => {
    const state = reduce([
      LLMEvent.textStart({ id: "t1" }),
@@ -172,7 +172,7 @@ describe("LLMClient tools", () => {
      expect(calls).toEqual([{ id: "call_projected", parameters: { prefix: "count" }, output: { count: "2" } }])
      expect(dispatched.result).toEqual({ type: "text", value: "count:2" })
      expect(dispatched.output).toEqual({ structured: { count: "2" }, content: [{ type: "text", text: "count:2" }] })
-      expect(dispatched.events).toMatchObject([
+      expect(dispatched.events).toEqual([
        LLMEvent.toolResult({
          id: "call_projected",
          name: "projected",
@@ -180,7 +180,6 @@ describe("LLMClient tools", () => {
          output: { structured: { count: "2" }, content: [{ type: "text", text: "count:2" }] },
        }),
      ])
-      expect(dispatched.events[0]?.itemId).toBeUndefined()
    }),
  )

@@ -198,7 +197,7 @@ describe("LLMClient tools", () => {
        LLMEvent.toolCall({ id: "call_1", name: "tool", input: {}, providerMetadata }),
      )

-      expect(dispatched.events).toMatchObject([
+      expect(dispatched.events).toEqual([
        LLMEvent.toolResult({
          id: "call_1",
          name: "tool",
@@ -207,13 +206,12 @@ describe("LLMClient tools", () => {
          providerMetadata,
        }),
      ])
-      expect(dispatched.events[0]?.itemId).toBeUndefined()

      const failed = yield* ToolRuntime.dispatch(
        {},
-        LLMEvent.toolCall({ id: "call_2", itemId: "fc_failed", name: "missing", input: {}, providerMetadata }),
+        LLMEvent.toolCall({ id: "call_2", name: "missing", input: {}, providerMetadata }),
      )
-      expect(failed.events).toMatchObject([
+      expect(failed.events).toEqual([
        LLMEvent.toolError({
          id: "call_2",
          name: "missing",
@@ -227,27 +225,6 @@ describe("LLMClient tools", () => {
          providerMetadata,
        }),
      ])
-      const errorItemID = failed.events.find(LLMEvent.is.toolError)?.itemId
-      const resultItemID = failed.events.find(LLMEvent.is.toolResult)?.itemId
-      expect(errorItemID).toBeUndefined()
-      expect(resultItemID).toBeUndefined()
-    }),
-  )
-
-  it.effect("does not derive a function output item id from the function call item id", () =>
-    Effect.gen(function* () {
-      const tool = Tool.make({
-        description: "Return text.",
-        parameters: Schema.Struct({}),
-        success: Schema.String,
-        execute: () => Effect.succeed("hello"),
-      })
-      const dispatched = yield* ToolRuntime.dispatch(
-        { tool },
-        LLMEvent.toolCall({ id: "call_1", itemId: "fc_existing", name: "tool", input: {} }),
-      )
-
-      expect(dispatched.events.find(LLMEvent.is.toolResult)?.itemId).toBeUndefined()
    }),
  )

@@ -460,7 +437,7 @@ describe("LLMClient tools", () => {
      )

      expect(dispatched.result).toEqual(callerOwned)
-      expect(dispatched.events).toMatchObject([
+      expect(dispatched.events).toEqual([
        LLMEvent.toolResult({
          id: "call_1",
          name: "eventful",
@@ -468,7 +445,6 @@ describe("LLMClient tools", () => {
          output: { structured: { ok: true }, content: [] },
        }),
      ])
-      expect(dispatched.events[0]?.itemId).toBeUndefined()
    }),
  )

@@ -18,6 +18,7 @@
  },
  "exports": {
    "./environment": "./src/environment/index.ts",
+    "./testing/environment-conformance": "./test/lib/environment-conformance.ts",
    "./session/runner": "./src/session/runner/index.ts",
    "./instructions": "./src/instructions/index.ts",
    "./*": "./src/*.ts"
@@ -116,11 +116,7 @@ export const createLLMEventPublisher = (bus: Pick<Bus.Interface, "publish">, inp
  })
  const currentAssistantMessageID = () =>
    stepStarted ? Effect.succeed(assistantMessageID) : Effect.die(new Error("Tool event before assistant step start"))
-  const providerState = (metadata: ProviderMetadata | undefined, itemId?: string) => {
-    const state = metadata?.[input.providerMetadataKey]
-    if (itemId === undefined) return state
-    return { ...(typeof state === "object" && state !== null && !Array.isArray(state) ? state : {}), itemId }
-  }
+  const providerState = (metadata: ProviderMetadata | undefined) => metadata?.[input.providerMetadataKey]
  const fragments = (
    name: string,
    ended: (id: string, value: string, ordinal: number, state?: Record<string, unknown>) => Effect.Effect<void>,
@@ -344,7 +340,7 @@ export const createLLMEventPublisher = (bus: Pick<Bus.Interface, "publish">, inp
        return
      case "text-start":
        outputStarted = true
-        const startedTextOrdinal = yield* text.start(event.id, providerState(event.providerMetadata, event.itemId))
+        const startedTextOrdinal = yield* text.start(event.id, providerState(event.providerMetadata))
        yield* bus.publish(SessionEvent.Text.Started, {
          sessionID: input.sessionID,
          assistantMessageID: yield* startAssistant(),
@@ -352,11 +348,7 @@ export const createLLMEventPublisher = (bus: Pick<Bus.Interface, "publish">, inp
        })
        return
      case "text-delta":
-        const deltaTextOrdinal = yield* text.append(
-          event.id,
-          event.text,
-          providerState(event.providerMetadata, event.itemId),
-        )
+        const deltaTextOrdinal = yield* text.append(event.id, event.text, providerState(event.providerMetadata))
        yield* bus.publish(SessionEvent.Text.Delta, {
          sessionID: input.sessionID,
          assistantMessageID: yield* currentAssistantMessageID(),
@@ -365,26 +357,23 @@ export const createLLMEventPublisher = (bus: Pick<Bus.Interface, "publish">, inp
        })
        return
      case "text-end":
-        yield* text.end(event.id, providerState(event.providerMetadata, event.itemId))
+        yield* text.end(event.id, providerState(event.providerMetadata))
        return
      case "reasoning-start":
        outputStarted = true
-        const startedReasoningOrdinal = yield* reasoning.start(
-          event.id,
-          providerState(event.providerMetadata, event.itemId),
-        )
+        const startedReasoningOrdinal = yield* reasoning.start(event.id, providerState(event.providerMetadata))
        yield* bus.publish(SessionEvent.Reasoning.Started, {
          sessionID: input.sessionID,
          assistantMessageID: yield* startAssistant(),
          ordinal: startedReasoningOrdinal,
-          state: providerState(event.providerMetadata, event.itemId),
+          state: providerState(event.providerMetadata),
        })
        return
      case "reasoning-delta":
        const deltaReasoningOrdinal = yield* reasoning.append(
          event.id,
          event.text,
-          providerState(event.providerMetadata, event.itemId),
+          providerState(event.providerMetadata),
        )
        yield* bus.publish(SessionEvent.Reasoning.Delta, {
          sessionID: input.sessionID,
@@ -394,7 +383,7 @@ export const createLLMEventPublisher = (bus: Pick<Bus.Interface, "publish">, inp
        })
        return
      case "reasoning-end":
-        yield* reasoning.end(event.id, providerState(event.providerMetadata, event.itemId))
+        yield* reasoning.end(event.id, providerState(event.providerMetadata))
        return
      case "tool-input-start":
        outputStarted = true
@@ -438,7 +427,7 @@ export const createLLMEventPublisher = (bus: Pick<Bus.Interface, "publish">, inp
          id: event.id,
          input: asRecord(event.input),
          executed: tool.providerExecuted,
-          state: providerState(event.providerMetadata, event.itemId),
+          state: providerState(event.providerMetadata),
        })
        return
      }
@@ -456,7 +445,7 @@ export const createLLMEventPublisher = (bus: Pick<Bus.Interface, "publish">, inp
        }
        tool.settled = true
        const executed = event.providerExecuted === true || tool.providerExecuted
-        const resultState = providerState(event.providerMetadata, event.itemId)
+        const resultState = providerState(event.providerMetadata)
        if (event.result.type === "error") {
          yield* bus.publish(SessionEvent.Tool.Failed, {
            sessionID: input.sessionID,
@@ -496,7 +485,7 @@ export const createLLMEventPublisher = (bus: Pick<Bus.Interface, "publish">, inp
              : { type: "tool.execution", message: event.message },
          ...failureSnapshot(tool),
          executed: tool.providerExecuted,
-          resultState: providerState(event.providerMetadata, event.itemId),
+          resultState: providerState(event.providerMetadata),
        })
        return
      }
@@ -521,7 +510,8 @@ export const createLLMEventPublisher = (bus: Pick<Bus.Interface, "publish">, inp

  const progress = Effect.fnUntraced(function* (id: string, update: Tool.Metadata) {
    const tool = tools.get(id)
-    if (!tool?.called || tool.settled) return yield* Effect.die(new Error(`Tool progress outside running call: ${id}`))
+    if (!tool?.called || tool.settled)
+      return yield* Effect.die(new Error(`Tool progress outside running call: ${id}`))
    tool.progress = update
    yield* bus.publish(SessionEvent.Tool.Progress, {
      sessionID: input.sessionID,
@@ -532,7 +522,11 @@ export const createLLMEventPublisher = (bus: Pick<Bus.Interface, "publish">, inp
  })

  /** Publishes one canonical terminal event for a locally executed tool call. */
-  const toolExecution = Effect.fnUntraced(function* (id: string, name: string, result: Tool.Result) {
+  const toolExecution = Effect.fnUntraced(function* (
+    id: string,
+    name: string,
+    result: Tool.Result,
+  ) {
    const tool = tools.get(id)
    if (!tool?.called) return yield* Effect.die(new Error(`Tool execution before call: ${id}`))
    if (tool.name !== name)
@@ -1,10 +1,4 @@
-import {
-  Message,
-  ToolCallPart,
-  ToolResultPart,
-  type ContentPart,
-  type ProviderMetadata,
-} from "@opencode-ai/ai"
+import { Message, ToolCallPart, ToolResultPart, type ContentPart, type ProviderMetadata } from "@opencode-ai/ai"
 import { Option, Schema } from "effect"
 import type { Model } from "../../model"
 import { SessionMessage } from "../message"
@@ -72,46 +66,27 @@ const providerMetadata = (
  state: Record<string, unknown> | undefined,
 ): ProviderMetadata | undefined => (state === undefined ? undefined : { [provider]: state })

-const responseItemID = (state: Record<string, unknown> | undefined) =>
-  typeof state?.itemId === "string" ? state.itemId : undefined
-
-const portableProviderState = (state: Record<string, unknown> | undefined) => {
-  if (state === undefined || !("itemId" in state)) return state
-  const { itemId: _itemId, ...portable } = state
-  return portable
-}
-
 const toolInput = (tool: SessionMessage.AssistantTool) =>
  tool.state.status === "streaming"
    ? Option.getOrElse(decodeToolInput(tool.state.input), () => tool.state.input)
    : tool.state.input

-const toolCall = (
-  tool: SessionMessage.AssistantTool,
-  itemId: string | undefined,
-  providerMetadata: ProviderMetadata | undefined,
-): ContentPart =>
+const toolCall = (tool: SessionMessage.AssistantTool, providerMetadata: ProviderMetadata | undefined): ContentPart =>
  ToolCallPart.make({
    id: tool.id,
-    ...(itemId === undefined ? {} : { itemId }),
    name: tool.name,
    input: toolInput(tool),
    providerExecuted: tool.executed,
    providerMetadata,
  })

-const toolResult = (
-  tool: SessionMessage.AssistantTool,
-  itemId: string | undefined,
-  providerMetadata: ProviderMetadata | undefined,
-) => {
+const toolResult = (tool: SessionMessage.AssistantTool, providerMetadata: ProviderMetadata | undefined) => {
  if (tool.state.status === "completed") {
    // TODO: Materialize remote and managed URIs before provider-history lowering.
    const content = tool.state.content
    const single = content.length === 1 ? content[0] : undefined
    return ToolResultPart.make({
      id: tool.id,
-      ...(itemId === undefined ? {} : { itemId }),
      name: tool.name,
      result:
        single?.type === "text"
@@ -124,7 +99,6 @@ const toolResult = (
  if (tool.state.status === "error") {
    return ToolResultPart.make({
      id: tool.id,
-      ...(itemId === undefined ? {} : { itemId }),
      name: tool.name,
      result: { error: tool.state.error, content: tool.state.content ?? [] },
      resultType: "error",
@@ -144,13 +118,7 @@ const assistant = (message: SessionMessage.Assistant, model: Model.Ref, provider
        {
          type: "text",
          text: item.text,
-          itemId: reuseProviderMetadata ? responseItemID(item.state) : undefined,
-          providerMetadata: sameProvider
-            ? providerMetadata(
-                providerMetadataKey,
-                reuseProviderMetadata ? item.state : portableProviderState(item.state),
-              )
-            : undefined,
+          providerMetadata: sameProvider ? providerMetadata(providerMetadataKey, item.state) : undefined,
        },
      ]
    if (item.type === "reasoning")
@@ -159,7 +127,6 @@ const assistant = (message: SessionMessage.Assistant, model: Model.Ref, provider
            {
              type: "reasoning",
              text: item.text,
-              itemId: responseItemID(item.state),
              providerMetadata: providerMetadata(providerMetadataKey, item.state),
            },
          ]
@@ -171,7 +138,6 @@ const assistant = (message: SessionMessage.Assistant, model: Model.Ref, provider
      (sameModel && item.executed === true && (item.state.status === "completed" || item.state.status === "error"))
    const call = toolCall(
      item,
-      reuseToolProviderMetadata ? responseItemID(item.providerState) : undefined,
      reuseToolProviderMetadata ? providerMetadata(providerMetadataKey, item.providerState) : undefined,
    )
    if (item.executed !== true) return [call]
@@ -179,11 +145,6 @@ const assistant = (message: SessionMessage.Assistant, model: Model.Ref, provider
    // replay must survive a model switch within the same provider.
    const result = toolResult(
      item,
-      reuseToolProviderMetadata
-        ? responseItemID(item.providerResultState ?? (item.executed === true ? item.providerState : undefined))
-        : sameProvider && item.executed === true
-          ? responseItemID(item.providerResultState)
-          : undefined,
      reuseToolProviderMetadata
        ? providerMetadata(providerMetadataKey, item.providerResultState ?? item.providerState)
        : sameProvider && item.executed === true && item.providerResultState !== undefined
@@ -202,8 +163,9 @@ const assistant = (message: SessionMessage.Assistant, model: Model.Ref, provider
    .map((item) =>
      toolResult(
        item,
-        responseItemID(item.providerResultState) ?? `fco_${item.id}`,
-        reuseProviderMetadata ? providerMetadata(providerMetadataKey, item.providerResultState) : undefined,
+        reuseProviderMetadata
+          ? providerMetadata(providerMetadataKey, item.providerResultState ?? item.providerState)
+          : undefined,
      ),
    )
    .filter((message) => message !== undefined)
@@ -242,7 +204,7 @@ function toLLMMessage(message: SessionMessage.Info, model: Model.Ref, providerMe
    case "skill":
      return [Message.make({ id: message.id, role: "user", content: message.text, metadata: message.metadata })]
    case "system":
-      return [Message.make({ id: message.id, role: "system", content: message.text })]
+      return [Message.system(message.text)]
    case "shell":
      return [
        Message.make({
@@ -118,12 +118,13 @@ const layer = Layer.effect(
        yield* hooks.trigger("tool", "execute.after", afterEvent)
        return yield* afterEvent.error
      }
+      const content = yield* normalizeImages(execution.value.content)
      const afterEvent: PluginHooks.Domains["tool"]["execute.after"] = {
        ...base,
        status: "completed",
        result: {
          ...(execution.value.output === undefined ? {} : { output: execution.value.output }),
-          content: execution.value.content,
+          content: content.length > 0 ? content : execution.value.content,
          ...(execution.value.metadata === undefined ? {} : { metadata: execution.value.metadata }),
        },
      }
@@ -110,13 +110,7 @@ describe("toLLMMessages", () => {
    )

    expect(messages.map((message) => message.role)).toEqual(["system", "user", "user", "user", "user"])
-    expect(messages[0]).toEqual(
-      Message.make({
-        id: "msg_system",
-        role: "system",
-        content: [{ type: "text", text: "Updated context\n\nOther context" }],
-      }),
-    )
+    expect(messages[0]).toEqual(Message.system("Updated context\n\nOther context"))
    expect(messages[1]).toEqual(
      Message.make({
        id: id("user"),
@@ -482,7 +476,6 @@ Recent work
      {
        type: "tool-result",
        id: "completed",
-        itemId: "fco_completed",
        name: "read",
        result: {
          type: "content",
@@ -520,7 +513,6 @@ Recent work
      {
        type: "reasoning",
        text: "Think",
-        itemId: "rs_1",
        providerMetadata: { provider: { itemId: "rs_1", reasoningEncryptedContent: "encrypted-state" } },
      },
    ])
@@ -610,7 +602,6 @@ Recent work
      {
        type: "tool-call",
        id: "hosted-completed",
-        itemId: "call_completed",
        name: "web_search",
        input: { query: "Effect" },
        providerExecuted: true,
@@ -619,7 +610,6 @@ Recent work
      {
        type: "tool-result",
        id: "hosted-completed",
-        itemId: "result_completed",
        name: "web_search",
        result: { type: "text", value: '{"found":true}' },
        providerExecuted: true,
@@ -630,7 +620,6 @@ Recent work
      {
        type: "tool-call",
        id: "hosted-failed",
-        itemId: "call_failed",
        name: "web_search",
        input: { query: "Effect" },
        providerExecuted: true,
@@ -639,7 +628,6 @@ Recent work
      {
        type: "tool-result",
        id: "hosted-failed",
-        itemId: "result_failed",
        name: "web_search",
        result: {
          type: "error",
@@ -710,7 +698,6 @@ Recent work
      {
        type: "tool-call",
        id: "hosted-old-model",
-        itemId: undefined,
        name: "web_search",
        input: { query: "Effect" },
        providerExecuted: true,
@@ -719,7 +706,6 @@ Recent work
      {
        type: "tool-result",
        id: "hosted-old-model",
-        itemId: "hosted-old-model",
        name: "web_search",
        result: { type: "text", value: '{"status":"completed"}' },
        providerExecuted: true,
@@ -732,7 +718,6 @@ Recent work
      {
        type: "tool-call",
        id: "local-old-model",
-        itemId: undefined,
        name: "read",
        input: { path: "README.md" },
        providerExecuted: false,
@@ -743,7 +728,6 @@ Recent work
      {
        type: "tool-result",
        id: "local-old-model",
-        itemId: "fco_local-old-model",
        name: "read",
        result: { type: "text", value: "Hello" },
        providerExecuted: false,
@@ -24,7 +24,9 @@ const capture = (providerMetadataKey = "anthropic", options?: { readonly interru
      const publish = Effect.sync(() => {
        const event = { id: Event.ID.create(), type: definition.type, data } as Event.Payload<typeof definition>
        published.push({
-          type: definition.durable ? Bus.versionedType(definition.type, definition.durable.version) : definition.type,
+          type: definition.durable
+            ? Bus.versionedType(definition.type, definition.durable.version)
+            : definition.type,
          data,
        })
        return event
@@ -64,10 +66,9 @@ const hostedResult = LLMEvent.toolResult({

 test("local tool success serializes media base64 once through canonical content", async () => {
  const { published, publisher } = capture()
-  const localCall = LLMEvent.toolCall({ ...call, itemId: "fc_call-image" })
-  await Effect.runPromise(publisher.publish(localCall))
+  await Effect.runPromise(publisher.publish(call))
  await Effect.runPromise(
-    publisher.toolExecution(localCall.id, localCall.name, {
+    publisher.toolExecution(call.id, call.name, {
      output: { type: "media", mime: "image/png" },
      content: [
        { type: "text", text: "Image read successfully" },
@@ -83,11 +84,6 @@ test("local tool success serializes media base64 once through canonical content"
  expect(success?.data).not.toHaveProperty("result")
  expect(success?.data).not.toHaveProperty("output")

-  const called = published.find((event) => event.type === "session.tool.called.1")?.data
-  expect(called).toMatchObject({ state: { itemId: "fc_call-image" } })
-  expect(success?.data).not.toHaveProperty("resultState")
-  expect(JSON.stringify(success?.data)).not.toContain('"itemId":"fc_call-image"')
-
  expect(success?.data).toMatchObject({
    content: [
      { type: "text", text: "Image read successfully" },
@@ -230,7 +226,9 @@ test("provider-executed tool metadata is flattened using the route key", async (
 test("binary failure emits no success event", async () => {
  const { published, publisher } = capture()
  await Effect.runPromise(publisher.publish(call))
-  await Effect.runPromise(publisher.failTool(call.id, { type: "tool.execution", message: "Cannot read binary file" }))
+  await Effect.runPromise(
+    publisher.failTool(call.id, { type: "tool.execution", message: "Cannot read binary file" }),
+  )
  expect(published.some((event) => event.type === "session.tool.success.2")).toBe(false)
  expect(published.some((event) => event.type === "session.tool.failed.2")).toBe(true)
 })
@@ -3,12 +3,10 @@ import { Agent } from "@opencode-ai/core/agent"
 import type { Permission } from "@opencode-ai/core/permission"
 import { AppNodeBuilder } from "@opencode-ai/core/effect/app-node-builder"
 import { Image } from "@opencode-ai/core/image"
-import { PluginHooks } from "@opencode-ai/core/plugin/hooks"
 import { Session } from "@opencode-ai/core/session"
 import { SessionMessage } from "@opencode-ai/core/session/message"
 import { Tool } from "@opencode-ai/core/tool"
 import type { Info } from "@opencode-ai/schema/tool"
-import { LayerNode } from "@opencode-ai/util/effect/layer-node"
 import { executeTool, toolDefinitions } from "./lib/tool"
 import { Cause, Deferred, Effect, Exit, Fiber, Layer, Option, Schema, SchemaGetter, SchemaIssue, Scope } from "effect"
 import { testEffect } from "./lib/effect"
@@ -28,14 +26,10 @@ const imageStore = Layer.mock(Image.Service, {
          maxBytes: 5,
        }),
      )
-    return Effect.succeed({
-      ...content,
-      content: Buffer.from(`${Buffer.from(content.content, "base64").toString()} normalized`).toString("base64"),
-      mime: "image/jpeg",
-    })
+    return Effect.succeed({ ...content, content: "bm9ybWFsaXplZA==", mime: "image/jpeg" })
  },
 })
-const registryLayer = AppNodeBuilder.build(LayerNode.group([Tool.node, PluginHooks.node]), [[Image.node, imageStore]])
+const registryLayer = AppNodeBuilder.build(Tool.node, [[Image.node, imageStore]])
 const it = testEffect(registryLayer)
 const identity = {
  agent: Agent.ID.make("build"),
@@ -350,7 +344,7 @@ describe("Tool", () => {
    }),
  )

-  it.effect("normalizes image tool output once and drops unresizable images", () =>
+  it.effect("normalizes image tool output at execution and drops unresizable images", () =>
    Effect.gen(function* () {
      const service = yield* Tool.Service
      yield* transform(service,
@@ -382,12 +376,7 @@ describe("Tool", () => {

      const execution = yield* executeTool(service, call("snapshot"))
      expect(execution.content).toEqual([
-        {
-          type: "file",
-          uri: "data:image/jpeg;base64,aW1hZ2Ugbm9ybWFsaXplZA==",
-          mime: "image/jpeg",
-          name: "frame.png",
-        },
+        { type: "file", uri: "data:image/jpeg;base64,bm9ybWFsaXplZA==", mime: "image/jpeg", name: "frame.png" },
        { type: "text", text: "snapshot" },
        { type: "text", text: "[1 image omitted: could not be decoded.]" },
        { type: "text", text: "[1 image omitted: could not be resized below the image size limit.]" },
@@ -395,34 +384,6 @@ describe("Tool", () => {
    }),
  )

-  it.effect("normalizes image content added by an after hook", () =>
-    Effect.gen(function* () {
-      const service = yield* Tool.Service
-      const hooks = yield* PluginHooks.Service
-      yield* transform(service, { hooked: constant("original") }, { codemode: false })
-      yield* hooks.register("tool", "execute.after", (event) =>
-        Effect.sync(() => {
-          if (event.status !== "completed") return
-          event.result = {
-            ...event.result,
-            content: [
-              { type: "file", uri: "data:image/png;base64,aW1hZ2U=", mime: "image/png", name: "hook.png" },
-            ],
-          }
-        }),
-      )
-
-      expect((yield* executeTool(service, call("hooked"))).content).toEqual([
-        {
-          type: "file",
-          uri: "data:image/jpeg;base64,aW1hZ2Ugbm9ybWFsaXplZA==",
-          mime: "image/jpeg",
-          name: "hook.png",
-        },
-      ])
-    }),
-  )
-
  it.effect("publishes progress metadata unchanged", () =>
    Effect.gen(function* () {
      const service = yield* Tool.Service
@@ -558,15 +558,6 @@ const messageTexts = (request: LLMRequest, role: "user" | "system") =>
 const userTexts = (request: LLMRequest) => messageTexts(request, "user")
 const systemTexts = (request: LLMRequest) => messageTexts(request, "system")
 const messageRoles = (request: LLMRequest | undefined) => request?.messages.map((message) => message.role)
-const withoutItemIDs = (messages: LLMRequest["messages"]) =>
-  messages.map((message) => ({
-    role: message.role,
-    content: message.content.map((part) => {
-      if (!("itemId" in part)) return part
-      const { itemId: _itemId, ...content } = part
-      return content
-    }),
-  }))

 const recordedEventTypes = (id: Session.ID) =>
  Effect.gen(function* () {
@@ -865,8 +856,8 @@ describe("SessionRunnerLLM", () => {
      yield* Fiber.join(renamed)

      expect(requests).toHaveLength(5)
-      expect(withoutItemIDs(requests[2]!.messages)).toContainEqual(withoutItemIDs([Message.user("First prompt")])[0])
-      expect(withoutItemIDs(requests[4]!.messages)).toContainEqual(withoutItemIDs([Message.user("First prompt")])[0])
+      expect(requests[2]?.messages).toContainEqual(Message.user("First prompt"))
+      expect(requests[4]?.messages).toContainEqual(Message.user("First prompt"))
      expect((yield* session.get(sessionID)).title).toBe("Generated title")
    }),
  )
@@ -891,7 +882,7 @@ describe("SessionRunnerLLM", () => {
      // A hook-removed call fails independently and continues while step allowance remains.
      expect(requests).toHaveLength(2)
      expect(requests[0]?.system.map((part) => part.text)).toEqual(["Hooked system"])
-      expect(withoutItemIDs(requests[0]!.messages)).toEqual(withoutItemIDs([Message.user("Hooked message")]))
+      expect(requests[0]?.messages).toEqual([Message.user("Hooked message")])
      expect(requests[0]?.tools.map((tool) => tool.name)).not.toContain("echo")
      expect(requests[0]?.tools.map((tool) => tool.name)).not.toContain("unregistered")
      expect(executions).toEqual([])
@@ -1316,14 +1307,12 @@ describe("SessionRunnerLLM", () => {
      systemBaseline = "Changed context"
      yield* runPrompt(session, "Second")

-      const firstSnapshot = PromptCacheDiagnostics.snapshot(requests[0]!)
-      const secondSnapshot = PromptCacheDiagnostics.snapshot(requests[1]!)
-      expect(PromptCacheDiagnostics.compare(firstSnapshot, secondSnapshot)).toEqual({
-        status: "append-only",
-        previousMessages: 1,
-        currentMessages: 3,
-      })
-      expect(secondSnapshot.messages[0]).toEqual(firstSnapshot.messages[0])
+      expect(
+        PromptCacheDiagnostics.compare(
+          PromptCacheDiagnostics.snapshot(requests[0]),
+          PromptCacheDiagnostics.snapshot(requests[1]),
+        ),
+      ).toEqual({ status: "append-only", previousMessages: 1, currentMessages: 3 })
      expect(requests.map((request) => request.system.map((part) => part.text))).toEqual([
        [defaultSystem, "Initial context"],
        [defaultSystem, "Initial context"],
@@ -2543,24 +2532,9 @@ describe("SessionRunnerLLM", () => {
        {
          type: "reasoning",
          text: "Encrypted thought",
-          itemId: "rs_1",
          providerMetadata: { openai: { itemId: "rs_1", reasoningEncryptedContent: "encrypted-state" } },
        },
      ])
-
-      yield* admit(session, "Continue again")
-      yield* TestLLM.push([])
-      yield* session.resume(sessionID)
-
-      expect(requests[2]?.messages[1]?.content.map((part) => ("itemId" in part ? part.itemId : undefined))).toEqual(
-        requests[1]?.messages[1]?.content.map((part) => ("itemId" in part ? part.itemId : undefined)),
-      )
-      expect(
-        PromptCacheDiagnostics.compare(
-          PromptCacheDiagnostics.snapshot(requests[1]!),
-          PromptCacheDiagnostics.snapshot(requests[2]!),
-        ),
-      ).toEqual({ status: "append-only", previousMessages: 3, currentMessages: 4 })
    }),
  )

@@ -2639,7 +2613,6 @@ describe("SessionRunnerLLM", () => {
        {
          type: "tool-call",
          id: "hosted-search",
-          itemId: "hosted-search",
          name: "web_search",
          input: { query: "Effect" },
          providerExecuted: true,
@@ -20,7 +20,8 @@
    "@opencode-ai/simulation": "workspace:*",
    "@opencode-ai/util": "workspace:*",
    "drizzle-orm": "catalog:",
-    "effect": "catalog:"
+    "effect": "catalog:",
+    "modal": "0.9.0"
  },
  "devDependencies": {
    "@tsconfig/bun": "catalog:",
@@ -0,0 +1,179 @@
+import { Effect, Sink, Stream } from "effect"
+import { systemError } from "effect/PlatformError"
+import type { Command, KillOptions } from "effect/unstable/process/ChildProcess"
+import { ExitCode, make, makeHandle, ProcessId } from "effect/unstable/process/ChildProcessSpawner"
+import type { Driver } from "@opencode-ai/core/environment"
+import type { ModalClientParams, Sandbox, SandboxCreateParams } from "modal"
+
+const INNER_WRAPPER = `
+pidfile=$1
+shift
+printf "%s" "$$" > "$pidfile"
+trap 'rm -f -- "$pidfile"' EXIT
+"$@"
+`
+
+const KILL = `
+i=0
+while [ ! -s "$1" ] && [ "$i" -lt 250 ]; do sleep 0.02; i=$((i + 1)); done
+if [ -s "$1" ]; then
+  pid=$(cat "$1")
+  /bin/kill "-$2" "-$pid" 2>/dev/null || true
+else
+  exit 47
+fi
+`
+
+export interface ModalImageSpec {
+  readonly registry: string
+  readonly dockerfileCommands: ReadonlyArray<string>
+}
+
+export interface ModalSandboxOptions {
+  readonly app: string
+  readonly client?: ModalClientParams
+  readonly image?: ModalImageSpec
+  readonly sandbox?: SandboxCreateParams
+}
+
+/**
+ * Ubuntu supplies the GNU coreutils and findutils required by the derived Files
+ * scripts. Busybox images do not satisfy the Environment contract.
+ */
+export const ubuntuImage: ModalImageSpec = {
+  registry: "ubuntu:24.04",
+  dockerfileCommands: [
+    "RUN apt-get update && apt-get install -y --no-install-recommends git bash ripgrep ca-certificates coreutils findutils util-linux",
+  ],
+}
+
+/** Creates a Modal sandbox lazily, keeping the SDK off the server startup path when Modal is unused. */
+export const createModalSandbox = async (options: ModalSandboxOptions) => {
+  const { ModalClient } = await import("modal")
+  const client = new ModalClient(options.client)
+  const app = await client.apps.fromName(options.app, { createIfMissing: true })
+  const imageSpec = options.image ?? ubuntuImage
+  const image = client.images.fromRegistry(imageSpec.registry).dockerfileCommands([...imageSpec.dockerfileCommands])
+  const sandbox = await client.sandboxes.create(app, image, options.sandbox)
+  return {
+    driver: makeModalDriver(sandbox),
+    sandbox,
+    terminate: () => sandbox.terminate(),
+  }
+}
+
+/**
+ * Adapts Modal exec to the Environment driver. Files intentionally has no native
+ * overrides: Modal exec and filesystem tools share the same roughly 175ms floor,
+ * so the derived exec defaults are the simplest implementation with no measured loss.
+ *
+ * Modal cannot signal a ContainerProcess. Each command therefore starts a new
+ * process group and records its leader in a unique pid file; kill runs a second
+ * sandbox command that signals that group. Pid files are removed best-effort.
+ */
+export const makeModalDriver = (sandbox: Sandbox): Driver => {
+  const spawn = Effect.fnUntraced(function* (command: Command) {
+    if (command._tag === "PipedCommand") {
+      return yield* Effect.fail(spawnError("spawn", "piped commands unsupported"))
+    }
+    if (command.options.additionalFds) {
+      return yield* Effect.fail(spawnError("spawn", "additional file descriptors unsupported"))
+    }
+
+    const pidFile = `/tmp/opencode-process-${crypto.randomUUID()}.pid`
+    const env = compact(command.options.env)
+    const isolatedEnv =
+      command.options.extendEnv === false || (!command.options.extendEnv && command.options.env !== undefined)
+    const argv = isolatedEnv ? ["env", "-i", ...Object.entries(env ?? {}).map(([key, value]) => `${key}=${value}`)] : []
+    const process = yield* Effect.tryPromise({
+      try: () =>
+        sandbox.exec(
+          ["setsid", "--wait", "sh", "-c", INNER_WRAPPER, "sh", pidFile, ...argv, command.command, ...command.args],
+          {
+            mode: "binary",
+            stdout: "pipe",
+            stderr: "pipe",
+            workdir: command.options.cwd,
+            env: isolatedEnv ? undefined : env,
+          },
+        ),
+      catch: (cause) => spawnError("spawn", undefined, cause),
+    })
+
+    const onError = (cause: unknown) => spawnError("process", undefined, cause)
+    let exited = false
+    const writer = process.stdin.getWriter()
+    let closingStdin: Promise<void> | undefined
+    const waited = process.wait().then((code) => {
+      exited = true
+      if (!closingStdin) writer.releaseLock()
+      return code
+    })
+    const exitCode = Effect.tryPromise({ try: () => waited, catch: onError }).pipe(Effect.map(ExitCode))
+    const kill = (options?: KillOptions) => {
+      if (exited) return Effect.void
+      return Effect.tryPromise({
+        try: async () => {
+          const killer = await sandbox.exec(["sh", "-c", KILL, "sh", pidFile, options?.killSignal ?? "SIGTERM"], {
+            stdout: "pipe",
+            stderr: "pipe",
+          })
+          const code = await killer.wait()
+          if (code !== 0) throw new Error(`modal kill exited ${code}`)
+        },
+        catch: onError,
+      }).pipe(Effect.andThen(exitCode), Effect.asVoid)
+    }
+
+    yield* Effect.addFinalizer(() => kill(command.options).pipe(Effect.ignore))
+
+    const closeStdin = Effect.tryPromise({
+      try: () => (closingStdin ??= writer.close().finally(() => writer.releaseLock())),
+      catch: onError,
+    })
+    const writeStdin = Sink.forEach((chunk: Uint8Array) =>
+      Effect.tryPromise({ try: () => writer.write(chunk), catch: onError }),
+    )
+    const inputConfig = command.options.stdin
+    const inputOptions =
+      inputConfig !== undefined && typeof inputConfig === "object" && !Stream.isStream(inputConfig)
+        ? inputConfig
+        : undefined
+    const input = inputOptions?.stream ?? inputConfig
+    const stdin = inputOptions?.endOnDone === false ? writeStdin : writeStdin.pipe(Sink.ensuring(closeStdin))
+    if (input === "ignore") {
+      yield* closeStdin
+    }
+    if (Stream.isStream(input)) {
+      yield* Effect.forkScoped(Stream.run(input, stdin))
+    }
+
+    const stdout = Stream.fromReadableStream({ evaluate: () => process.stdout, onError })
+    const stderr = Stream.fromReadableStream({ evaluate: () => process.stderr, onError })
+    return makeHandle({
+      pid: ProcessId(crypto.getRandomValues(new Uint32Array(1))[0]),
+      exitCode,
+      isRunning: Effect.sync(() => !exited),
+      kill,
+      stdin,
+      stdout,
+      stderr,
+      all: Stream.merge(stdout, stderr),
+      getInputFd: () => Sink.fail(spawnError("getInputFd", "unsupported")),
+      getOutputFd: () => Stream.fail(spawnError("getOutputFd", "unsupported")),
+      unref: Effect.succeed(Effect.void),
+    })
+  })
+
+  return { spawner: make(spawn) }
+}
+
+const compact = (env: Record<string, string | undefined> | undefined) => {
+  if (!env) return undefined
+  return Object.fromEntries(Object.entries(env).flatMap(([key, value]) => (value === undefined ? [] : [[key, value]])))
+}
+
+const spawnError = (method: string, description?: string, cause?: unknown) =>
+  systemError({ _tag: "Unknown", module: "ModalDriver", method, description, cause })
+
+export * as ModalDriver from "./modal"
@@ -0,0 +1,67 @@
+import fs from "node:fs"
+import os from "node:os"
+import path from "node:path"
+import { afterAll, expect, test } from "bun:test"
+import { Effect } from "effect"
+import { ChildProcess } from "effect/unstable/process"
+import { Failed, makeFiles } from "@opencode-ai/core/environment"
+import { environmentConformance } from "@opencode-ai/core/testing/environment-conformance"
+import { createModalSandbox } from "../src/workspace/modal"
+
+const enabled =
+  !!process.env.OPENCODE_TEST_MODAL &&
+  ((!!process.env.MODAL_TOKEN_ID && !!process.env.MODAL_TOKEN_SECRET) ||
+    fs.existsSync(path.join(os.homedir(), ".modal.toml")))
+const modalTest = enabled ? test : test.skip
+const root = `/tmp/opencode-environment-${crypto.randomUUID()}`
+const sandbox = enabled
+  ? createModalSandbox({
+      app: "opencode-environment-tests",
+      sandbox: { timeoutMs: 10 * 60 * 1000 },
+    })
+  : undefined
+
+modalTest(
+  "kills a running modal process",
+  async () => {
+    const value = await sandbox!
+    const started = performance.now()
+    const exitCode = await Effect.runPromise(
+      Effect.scoped(
+        Effect.gen(function* () {
+          const handle = yield* value.driver.spawner.spawn(ChildProcess.make("sleep", ["300"]))
+          yield* handle.kill()
+          return yield* handle.exitCode
+        }),
+      ).pipe(Effect.timeout("10 seconds")),
+    )
+    expect(exitCode).not.toBe(0)
+    expect(performance.now() - started).toBeLessThan(10_000)
+  },
+  15_000,
+)
+
+environmentConformance(
+  "modal environment",
+  () =>
+    Effect.promise(async () => {
+      const value = await sandbox!
+      return {
+        files: makeFiles(value.driver),
+        root,
+        symlink: (target, link) =>
+          Effect.scoped(value.driver.spawner.exitCode(ChildProcess.make("ln", ["-s", "--", target, link]))).pipe(
+            Effect.flatMap((code) =>
+              code === 0 ? Effect.void : Effect.fail(new Failed({ path: link, cause: new Error(`ln exited ${code}`) })),
+            ),
+            Effect.mapError((cause) => (cause instanceof Failed ? cause : new Failed({ path: link, cause }))),
+          ),
+      }
+    }),
+  !enabled,
+)
+
+afterAll(async () => {
+  if (!sandbox) return
+  await (await sandbox).terminate()
+})
Author	SHA1	Message	Date
Kit Langton	e5caf4902d	refactor(server): simplify modal process lifecycle	2026-08-07 13:21:44 -04:00
Kit Langton	5ff39f379c	feat(server): add modal environment driver	2026-08-07 13:16:08 -04:00
Kit Langton	255d35009b	feat(core): export environment conformance suite	2026-08-07 13:16:05 -04:00