AgentProbe/docs/INDEX.md

docs Index

Purpose

Canonical repository knowledge base for agents and humans. This directory holds the long-lived, versioned sources of truth for architecture, product behavior, engineering standards, plans, references, and generated inventories.

File conventions

• Markdown files only.
• Top-level files define cross-cutting rules, contracts, and navigation.
• Topic-specific knowledge belongs in subdirectories with their own INDEX.md.
• Generated artifacts live only under generated/ and are refreshed by script.
• If knowledge is important for future agent runs, encode it here instead of leaving it in chat, tickets, or memory.

Files

• README.md — Human and agent navigation guide for the docs tree
• ARCHITECTURE.md — Layered TypeScript system map and dependency direction
• DESIGN.md — Agent-first design principles and repository norms
• HARNESS.md — PR, validation, and operating contract
• PLANS.md — Planning workflow, decision logging, and debt tracking
• PRODUCT_SENSE.md — Product intent and CLI UX priorities
• QUALITY_SCORE.md — Generated quality grade and cleanup pressure
• RELIABILITY.md — Observability, latency budgets, and debugging rules
• SECURITY.md — Security posture, boundaries, and secret handling

Subdirectories

• design-docs/INDEX.md — Durable design decisions and core beliefs
• exec-plans/INDEX.md — Active plans, completed plans, and debt tracker
• generated/INDEX.md — Script-owned generated documentation
• playbooks/INDEX.md — Repeatable operational runbooks
• product-specs/INDEX.md — Product behavior specs and coverage snapshots
• references/INDEX.md — Toolchain, encoding, observability, and quality references

• ARCHITECTURE.md
• DESIGN.md
• HARNESS.md
• PLANS.md
• PRODUCT_SENSE.md
• QUALITY_SCORE.md
• README.md
• RELIABILITY.md
• SECURITY.md

• design-docs/INDEX.md
• exec-plans/INDEX.md
• generated/INDEX.md
• playbooks/INDEX.md
• product-specs/INDEX.md
• references/INDEX.md