address-reviews

Pass

Audited by Gen Agent Trust Hub on Mar 3, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it ingests and processes review feedback provided by external users which could contain malicious instructions.\n
  • Ingestion points: Data is gathered from GitHub PR reviews, inline comments, and conversation comments via gh api calls in Step 2.\n
  • Boundary markers: No specific boundary markers or instructions to ignore embedded commands are present in the skill definition.\n
  • Capability inventory: The skill has access to Bash (for git and gh), Read, Grep, and Glob tools, and can enter a planning mode to modify source code.\n
  • Sanitization: The skill does not perform sanitization or validation of the fetched comment content before processing it to create an actionable plan.\n- [COMMAND_EXECUTION]: The skill executes git and gh commands to retrieve repository metadata and PR details. While restricted to a specific allowed-tools list, these commands represent the primary interface for data retrieval and the execution surface.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 3, 2026, 01:37 AM