github-workflow-automation

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 0.90). The content contains intentional data-exfiltration and remote-modification behaviors (sending PR diffs/commit logs/context to external AI services like Anthropic), plus automated flows that can rewrite repository history (force-push on auto-rebase) and programmatically apply AI-provided code resolutions—together these are high-risk abuse vectors for leaking sensitive code/credentials and enabling remote modifications.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.90). The skill's workflows explicitly send untrusted, user-generated GitHub content (e.g., PR diffs via the "Get diff" + "AI Review" action in references/workflow_examples.md, issue title/body in .github/workflows/issue-triage.yml, and PR/issue context in the mention-bot.yml on-demand flow) to an LLM and then use the model's output to post reviews, apply labels, gate deployments, or perform git operations, so third-party content can materially influence agent actions.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.70). The workflows use the external GitHub Action actions/github-script@v7 (https://github.com/actions/github-script) which is fetched and executed at runtime and runs arbitrary JavaScript (including calling AI APIs and making repo changes), so it is a runtime external dependency that executes remote code and the skill depends on it.