Skip to content

docs: add MAI-1 agent contract to AGENTS.md#649

Open
javierfajardo85-rgb wants to merge 1 commit into
theskumar:mainfrom
javierfajardo85-rgb:mai1-inline-contract
Open

docs: add MAI-1 agent contract to AGENTS.md#649
javierfajardo85-rgb wants to merge 1 commit into
theskumar:mainfrom
javierfajardo85-rgb:mai1-inline-contract

Conversation

@javierfajardo85-rgb
Copy link
Copy Markdown

Adds an AGENTS.md containing a self-contained MAI-1 tool contract.

Why this helps: AI coding agents (Claude Code, GitHub Copilot, Cursor, Gemini CLI) can read the contract block to discover how to install and invoke this library without parsing the full README or hallucinating the API.

The block follows the MAI-1 four-section structure:

{
  "aid": "python-dotenv-v1",
  "logic":  { "input_schema": {...},  "output_schema": {...} },
  "trust":  { "reliability_score": 0.0, "latency_ms": 0       },
  "action": { "install_cmd": "...",     "execute_cmd": "..."   }
}

No external services are queried. The file lives entirely in this repo and can be maintained by the team like any other documentation.

Feel free to close if this doesn't fit the project.

javierfajardo85-rgb pushed a commit to javierfajardo85-rgb/aiaam-xyz that referenced this pull request May 24, 2026
… scorer

sanitizer (main.py):
- _sanitize_query now also strips literal \n \r sequences (two chars)
  that some agents send instead of real newline characters — fixes the
  'send transactional emailnnresponse' zero-result bug

dedup migration (main.py):
- _migrate_dedup_categories(): renames 'payment' → 'payments' in
  compiled_apis at startup; idempotent, runs on Railway deploy

mai_api.html:
- Added missing category options: finance, google, social
- Sorted alphabetically, matches backend validator exactly

reliability_scorer.py:
- Extended from GitHub-only to all 4 platforms:
  · PyPI: pypistats.org API (downloads/month → score 0.72–0.97)
  · npm: npmjs.org downloads API (same scale)
  · HuggingFace: HF API (downloads + likes → score 0.74–0.95)
- --platform flag to score one platform at a time
- All scorers share compute_pypi_score scale for consistency

B3 PRs:
- python-dotenv: theskumar/python-dotenv#649 ✅
- letta: blocked (org-level permissions, 404 on PR endpoint) — deferred

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
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.

1 participant