comet-github
Pass
Audited by Gen Agent Trust Hub on Aug 14, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is comprised entirely of instructional routing logic and a security-focused maintainer contract. It contains no executable scripts, remote dependencies, or automated write operations.
- [PROMPT_INJECTION]: No attempts to bypass safety filters, extract system prompts, or override agent instructions were detected. The skill uses clear, intent-based routing without adversarial language.
- [DATA_EXFILTRATION]: There are no network requests, hardcoded credentials, or sensitive file access patterns. The skill explicitly enforces a 'default read-only' policy for all maintenance activities.
- [INDIRECT_PROMPT_INJECTION]: The skill establishes robust boundaries for handling external data from GitHub by requiring independent verification and real-time status refreshes.
- Ingestion points: User-provided GitHub Issue/PR numbers and URLs (SKILL.md).
- Boundary markers: Instructions to separate evidence from suggestions and to refresh GitHub state before acting (SKILL.md).
- Capability inventory: Read access to local Git status, CI logs, and GitHub metadata (references/maintainer-contract.md).
- Sanitization: Mandatory cross-referencing of external data against actual source code and runtime behavior (references/maintainer-contract.md).
Audit Metadata