receiving-code-review

Pass

Audited by Gen Agent Trust Hub on Jul 3, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill provides strict behavioral instructions that override the agent's default interaction style, specifically forbidding expressions of gratitude or social agreement ("NEVER: 'You're absolutely right!'") to maintain a focused technical persona.
  • [PROMPT_INJECTION]: The skill is designed to process untrusted data in the form of external code review feedback, which constitutes an indirect prompt injection surface.
  • Ingestion points: External review comments from GitHub or other sources (referenced in SKILL.md).
  • Boundary markers: The instructions include explicit verification steps such as checking if suggestions "break existing functionality" or are "technically correct for THIS codebase."
  • Capability inventory: The skill references the ability to search the codebase using grep and perform network operations via the GitHub CLI (gh api).
  • Sanitization: No automated sanitization is mentioned; the skill relies on the agent's logical evaluation and manual verification steps.
  • [COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands, specifically grep, to search the codebase for feature usage as part of "YAGNI" (You Ain't Gonna Need It) checks.
  • [COMMAND_EXECUTION]: The skill directs the agent to use the GitHub CLI tool (gh api) to perform authenticated network requests for replying to specific pull request comments.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 3, 2026, 12:42 PM
Security Audit — agent-trust-hub — receiving-code-review