diff --git a/AGENTS.md b/AGENTS.md index b32a45f..88901a2 100644 --- a/AGENTS.md +++ b/AGENTS.md @@ -4,18 +4,26 @@ This project is designed for AI-first development. All agents MUST follow these protocols: +### 0. Context Loading (ALWAYS FIRST) +- **RULE**: ALWAYS read project context before starting ANY task. +- **PROCESS**: + 1. Read `docs/ARCHITECTURE.md` to understand system design. + 2. Read `docs/learnings/*.md` to learn from past corrections. + 3. Check `.agents/skills/` for relevant technology skills. +- **WHY**: These files contain critical knowledge from past work. Skipping them leads to repeated mistakes and inconsistent code. + ### 1. Technology Skill Acquisition - **RULE**: NEVER work on a technology without verified skills. - **PROCESS**: 1. If a task involves a new technology/library: **STOP**. - 2. Perform a `web_search` or `librarian` task to investigate best practices, common pitfalls, and API usage. - 3. Document findings in `.agents/skills/.md`. + 2. Perform a web search or use the dedicated task agent to investigate best practices, common pitfalls, and API usage. + 3. Document findings in `.agents/skills//SKILL.md`. 4. Only proceed with implementation once the skill file exists and is reviewed. ### 2. Architecture First - **RULE**: Architecture MUST be adapted/reviewed before coding starts. - **PROCESS**: - 1. For any new feature/package, update `docs/architecture/DECISIONS.md`. + 1. For any new feature/package, update `docs/DECISIONS.md`. 2. Ensure the architecture aligns with the overall project goals. ### 3. Learning from Corrections @@ -47,14 +55,13 @@ This project is designed for AI-first development. All agents MUST follow these 4. Update all affected documentation before marking task as complete. ## PROJECT OVERVIEW -- **Goal**: LLM-optimized AI project. - **Core Stack**: Python + FastAPI + uv (build system) ## WHERE TO LOOK -- `docs/architecture/`: System design and decisions. -- `docs/skills/`: Tech-specific knowledge base. +- `docs/ARCHITECTURE.md`: System design documentation. +- `docs/DECISIONS.md`: Architecture decisions. - `docs/learnings/`: Persistent memory of user corrections. -- `docs/tests/`: Global test configurations and results. +- `.agents/skills/`: Tech-specific knowledge base. - `src/llmock/`: Main application source code. - `tests/`: Test files.