run-health-review
Pass
Audited by Gen Agent Trust Hub on Nov 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data from GitHub environments, which could contain malicious instructions designed to influence the agent's actions during health reviews.
- Ingestion points: The skill ingests data from GitHub PR states, commit messages, and automation log lines to classify 'stalls' and monitor health signals (SKILL.md, Step 2 and Taxonomy sections).
- Boundary markers: Absent. There are no instructions to wrap processed data in delimiters or specifically ignore instructions embedded in the logs or PR descriptions.
- Capability inventory: The agent has the ability to post comments to PRs, assign workers to branches, and propose updates to other automations or prompt files based on its analysis (SKILL.md, Takeover protocol and Outputs section).
- Sanitization: Absent. The instructions do not specify any sanitization or validation of the external content before it is used for classification or reporting tasks.
Audit Metadata