vlad-ko/Claude Mate
Run an autonomous Claude-powered maintenance agent (docs, tests, dead-code, logic, security) on your repo.
View on GitHubTrust Signals
- Scorecard Score
- not yet scored
- Maintenance Recency
- Stale
- License
- None
Inputs
| name | description | required | default |
|---|---|---|---|
| mate | Which mate to run. One of: docs, tests, dead-code, logic, security. See the repo README for the full list. | yes | — |
| api-key | Anthropic API key (the CLAUDE_MATES_API_KEY secret). Passed to the Claude CLI (generic mates) or the Anthropic security scanner (security mate) as ANTHROPIC_API_KEY / claude-api-key. | yes | — |
| config-path | Path to .claude-mates.yml inside the consumer workspace. Used for per-mate enable/disable, model overrides, scope, and deny rules. | no | .claude-mates.yml |
| claude-cli-version | Version of @anthropic-ai/claude-code to install (drift mates only; ignored for security). | no | 2.1.97 |
Outputs
| name | description |
|---|---|
| outcome | What the mate produced on this run. One of: none (clean/error/empty/violations-only), findings (drift mates: Claude reported findings but made no file edits, rendered to the Job Summary; security mate: scanner found one or more security findings, posted as inline PR comments), pr (drift mate: a PR was opened — check pr-url for success). |
| status | Classification of Claude's output (drift mates). Security mate always emits ok when the scanner completes. |
| issue-url | URL of a GitHub issue linked to this run. Always empty: drift mates no longer auto-file issues (findings go to the Job Summary), and the security mate posts inline PR comments rather than issues. |
| pr-url | URL of the GitHub PR created by this run (drift mates only; empty for security mate). |
| findings-count | Number of findings reported (security mate only; empty for other mates). |