pr-address-bot-reviews

Pass

Audited by Gen Agent Trust Hub on Mar 1, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it is designed to act upon feedback from external bots. \n
  • Ingestion points: The agent fetches bot-authored reviews and comments using the 'gh' CLI and GitHub API. \n
  • Boundary markers: There are no explicit delimiters used to separate untrusted bot content from the agent's primary instructions. \n
  • Capability inventory: The agent can modify the filesystem, commit/push code, and run shell commands for testing. \n
  • Sanitization: The skill does not sanitize or validate external feedback before processing. \n- [COMMAND_EXECUTION]: The agent is instructed to implement fixes and run tests based on bot comments, creating a risk that malicious suggestions could lead to the execution of harmful shell commands. \n- [DATA_EXFILTRATION]: The 'scripts/pr-thread-reply' script uses the '--body-file' argument to read local files and post them to the PR. An attacker could potentially trick the agent into reading sensitive files and posting their contents publicly.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 1, 2026, 12:37 AM