skills/tmchow/agent-skills/clawpatch/Gen Agent Trust Hub

clawpatch

Pass

Audited by Gen Agent Trust Hub on May 20, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill provides instructions for the agent to execute various terminal commands, including clawpatch, npm, jq, and git. These are used for tool initialization, dependency management, output parsing, and repository state management.
  • [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface as it directs the agent to ingest and parse JSON findings generated by the clawpatch CLI from the analyzed repository.
  • Ingestion points: Findings are ingested via clawpatch report --json and processed using jq (SKILL.md).
  • Boundary markers: No explicit boundary markers or isolation instructions are provided for handling the ingested findings.
  • Capability inventory: The skill uses terminal access for executing CLI tools and performs file system operations such as modifying .gitignore and creating temporary files (SKILL.md).
  • Sanitization: No specific sanitization or validation steps are defined for the data extracted from the CLI tool's findings.
Audit Metadata
Risk Level
SAFE
Analyzed
May 20, 2026, 06:28 AM
Security Audit — agent-trust-hub — clawpatch