review

Pass

Audited by Gen Agent Trust Hub on Mar 7, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection as it ingests and processes untrusted pull request comments and review summaries from external contributors.
  • Ingestion points: Pull request comments and review bodies are fetched from the GitHub API using gh api in SKILL.md (Step 1).
  • Boundary markers: The skill does not define clear delimiters or instructions to the agent to treat fetched comment content as untrusted data or to ignore embedded commands.
  • Capability inventory: The skill possesses the ability to execute shell commands (gh), read local source files, modify files using the Edit tool, and post content back to GitHub.
  • Sanitization: There is no evidence of sanitization, filtering, or escaping of the fetched GitHub comment strings before they are presented to the agent for evaluation.
  • [COMMAND_EXECUTION]: The skill relies on the GitHub CLI (gh) to perform its primary functions, including querying pull request status and posting replies. While these are necessary for the skill's purpose, they provide a vector for command-line interaction that could be leveraged if the agent's logic is subverted by malicious input.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 7, 2026, 08:22 AM