implement-task
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from multiple sources which could contain hidden instructions intended to override agent behavior. \n
- Ingestion points: Processes developer-provided task statements (SKILL.md), external Figma mockups via MCP (SKILL.md), and project-specific or user-defined standards files (references/standards.md). \n
- Boundary markers: No specific delimiters or "ignore embedded instructions" warnings are defined for the ingested data. \n
- Capability inventory: Performs shell command execution (ocr CLI, run skill), file writes (.claude/verify.md), and git operations (git commit). \n
- Sanitization: External content is interpolated into the workflow without explicit validation or escaping. \n- [COMMAND_EXECUTION]: The skill invokes the ocr (Open Code Review) CLI tool in delegation mode to preview files and apply review rules, passing file paths as arguments to these shell commands (SKILL.md). \n- [EXTERNAL_DOWNLOADS]: The skill metadata includes an update command 'npx skills update' which facilitates remote code execution from the package registry for maintenance purposes (SKILL.md).
Audit Metadata