How I built a GitHub Actions connector for Cognee — and what it can do for your AI agents

Description and use

A GitHub Actions connector for Cognee was built to ingest workflow definitions, run attempts, and job failures into an AI memory layer, enabling AI queries over CI history.

Problem

Your CI history lives in GitHub but your AI agent can’t query it.

Outcome

The connector ingests workflow definitions, run attempts, and job details, caching unchanged files and supporting incremental sync. It limits results to 1,000 per query and handles deletions via tombstones.

Workflow

• 1. Install the connector package.
• 2. Create a source with repositories and a GitHub token: source = github_actions_source(repositories=["owner/repo"], token="...").
• 3. Ingest data into Cognee: await cognee.remember(source, dataset_name="github_actions").
• 4. Query the ingested data: answer = await cognee.search("Which workflows failed on main this week?")

Evidence

Adds a GitHub Actions data-source connector for cognee-community, following the Notion connector conventions.

Source

Workflow definitions — triggers, job names, step names, uses: action references. Raw shell scripts and env/secrets are excluded.

Source

All tests use a fully mocked FakeGitHubClient — no live token required.

Source