4-step-program

Pass

Audited by Gen Agent Trust Hub on Feb 27, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill uses authoritative language (e.g., 'MANDATORY', 'CRITICAL', '100% coverage') to enforce a strict workflow. While intended for quality control, these patterns represent the agent's internal steering logic.
  • [INDIRECT_PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection by ingesting untrusted external data.
  • Ingestion points: The skill explicitly instructs the agent to read external content using gh issue view <number> and to analyze pull request changes (SKILL.md).
  • Boundary markers: There are no instructions for the agent to use delimiters or 'ignore' tags when processing the body of GitHub issues or PR comments.
  • Capability inventory: The agent has the capability to delegate tasks (assign_task), communicate with other agents (send_message_to_agent), and write content back to GitHub (mcp__github__create_pull_request_review).
  • Sanitization: No sanitization or validation logic is specified for the data retrieved from GitHub before it is used to formulate delegation prompts or review comments.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 27, 2026, 08:48 AM