Skip to content

chore: add AI agent context files (AGENTS.md, USERS.md, conventions) - #459

Open
e-lo wants to merge 3 commits into
mainfrom
claude/agitated-wescoff
Open

chore: add AI agent context files (AGENTS.md, USERS.md, conventions)#459
e-lo wants to merge 3 commits into
mainfrom
claude/agitated-wescoff

Conversation

@e-lo

@e-lo e-lo commented Apr 13, 2026

Copy link
Copy Markdown
Collaborator

Summary

  • Adds tool-agnostic AGENTS.md (~400 tokens) so any AI coding assistant (Claude Code, Cursor, Copilot, Aider, Windsurf) gets consistent project context
  • Adds CLAUDE.md as a thin Claude-specific delegate to AGENTS.md
  • Adds USERS.md with 4 user personas (transportation planners, model developers, tool developers, CI/automation) and UX principles
  • Adds docs/conventions.md covering coding style, validation strategy (when/where to run pandera and FK checks), pandas/pandera/geopandas patterns, testing, and commits
  • Extends docs/design.md with 7 mermaid diagrams: ecosystem data flow, core object model, roadway selection, project application flow, scoped property resolution, IO pipeline
  • Updates CONTRIBUTING.md with diagram maintenance convention and expanded testing section

Closes #458

Design Decisions

  • Agent-agnostic: all project knowledge in AGENTS.md + docs/. No tool-specific plugins committed to repo.
  • No duplication: AGENTS.md is a concise index pointing to docs/conventions.md, docs/design.md, CONTRIBUTING.md, USERS.md. Details live in one place only.
  • Validation strategy: documents when pandera/FK validation should and should not run (IO boundaries and project boundaries, not between internal functions) per performance requirements.
  • No nested AGENTS.md: mermaid diagrams in docs/design.md replace the need for per-directory context files.

Test plan

  • Verify mkdocs serve renders new docs/conventions.md and mermaid diagrams in docs/design.md correctly
  • Confirm no content duplication between AGENTS.md and docs/conventions.md
  • Review persona descriptions in USERS.md for accuracy
  • Check that mermaid class diagrams match actual class structure

🤖 Generated with Claude Code

e-lo and others added 2 commits April 13, 2026 14:32
…d development

Implements #458. Adds tool-agnostic context
files so any AI coding assistant gets consistent project context:

- AGENTS.md: token-efficient root context with doc pointers and ecosystem overview
- CLAUDE.md: thin Claude-specific file delegating to AGENTS.md
- USERS.md: 4 user personas and UX principles for design decisions
- docs/conventions.md: coding style, validation strategy, testing, commits
- docs/design.md: 7 mermaid diagrams (object model, project flow, ecosystem, etc.)
- CONTRIBUTING.md: diagram maintenance convention, expanded testing section

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
@github-actions

Copy link
Copy Markdown
Contributor

📚 Documentation has been built and deployed!

View the docs for this PR: https://network-wrangler.github.io/network_wrangler/claude/agitated-wescoff/

…e design

Adds three Anthropic Agent Skill-format guides at skills/ for AI coding
assistants: root-cause debugging discipline, measure-first performance
work, and user-persona + ecosystem-aware feature design. Readable by any
skill-aware agent (Claude Code, Codex) and by humans directly.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
@e-lo
e-lo requested a review from lmz April 14, 2026 19:30
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

🧹 Add AGENTS.md context file structure for AI coding assistants

1 participant