Repository readiness evidence for AI coding agents

This Actor checks a small, explicit handoff checklist for a public GitHub repository. It returns source links, byte counts, line counts, SHA-256 hashes, and literal required-link checks as structured dataset records.

What it does

Give it a public repository URL, a branch/tag/commit, and the paths that an agent must be able to inspect. It fetches the Git tree, reads only the requested files within the configured size limit, and writes:

• one summary record with ready or incomplete status;
• one path record for each required path;
• one link record for each required link.

The Actor does not run tests, execute repository code, scan private repositories, or claim that the code is correct. A present record proves that a particular public file was available at the selected ref and gives the agent a source URL and content hash. A separate CI or local test run is still required for behavioral correctness.

Example input

{
  "repository_url": "https://github.com/Abood991B/apify-content-program-actors",
  "ref": "main",
  "required_paths": [
    "README.md",
    "repository-readiness-actor/src/readiness.py",
    "repository-readiness-actor/tests/test_readiness.py"
  ],
  "required_links": [
    "https://docs.apify.com/integrations/mcp"
  ],
  "max_file_bytes": 1000000,
  "include_content_preview": false
}

Why an AI agent can use it

The output is intentionally boring: a checklist result, not a generated review. An agent can decide what to do next from explicit fields and follow the sourceUrl or rawUrl without guessing which branch or file was inspected. The original request is also kept in the default key-value store under REQUEST, and the summary is stored under SUMMARY.

Limits and responsible use

The implementation calls GitHub's public REST and raw-content endpoints. It does not bypass authentication. Respect GitHub's terms, rate limits, and the repository's license. The tree endpoint can report repository_tree_truncated; in that case an incomplete result is safer than treating an absent path as proof that a file does not exist.

Local development

python -m venv .venv
.\.venv\Scripts\Activate.ps1
python -m pip install -r requirements.txt
python -m unittest discover -s tests -v

The Actor is deployed from the repository root with the Apify CLI. The .actor folder contains the input and output schemas used by Apify Console and the Apify MCP server.