code-review-excellence
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted external data during pull request reviews, creating a standard surface for indirect prompt injection.\n
- Ingestion points: The skill instructions in
SKILL.mdinvolve reading pull request descriptions, code changes, and linked issues.\n - Boundary markers: There are no specific instructions or delimiters provided to the agent to treat the code being reviewed as data instead of potentially malicious instructions.\n
- Capability inventory: The skill references
scripts/pr-analyzer.py, which is intended to analyze pull request complexity and metadata.\n - Sanitization: While the skill encourages security analysis of the reviewed code, it does not specify sanitization or filtering protocols for the agent's own input processing of that code.
Audit Metadata