harden-branch

Pass

Audited by Gen Agent Trust Hub on Jul 8, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill documentation provides instructions to install required child skills from external GitHub repositories using the npx skills utility. Specifically, it references repositories belonging to the vendor (patinaproject/skills) and a third-party developer (mattpocock/skills).
  • Evidence: Shell commands in the "Required Child Skills" section.
  • Mitigation: The suggested installation commands include npm_config_ignore_scripts=true, which is a security best practice that prevents the execution of potentially malicious post-install scripts.
  • [PROMPT_INJECTION]: The skill functions as a coordinator that processes results from an automated reviewer and routes findings to implementation tools, which represents an indirect prompt injection surface.
  • Ingestion points: Output and findings generated by the review-branch skill (SKILL.md).
  • Boundary markers: The skill does not explicitly define delimiters or specific "ignore" instructions for the data ingested from the reviewer findings.
  • Capability inventory: The skill can indirectly trigger code changes and bug fixes by invoking the implement and diagnosing-bugs skills.
  • Sanitization: No explicit sanitization or validation of the findings' content is described before routing to the implementation phase.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 8, 2026, 08:20 AM
Security Audit — agent-trust-hub — harden-branch