risk-assess

Pass

Audited by Gen Agent Trust Hub on Aug 24, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The analysis script scripts/risk-analyze.js utilizes git and grep via a subprocess execution wrapper to gather metrics on code changes, blast radius, and churn. These operations are restricted to the local repository and are appropriate for a code analysis tool.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests untrusted code from the repository and instructs the agent to provide a qualitative interpretation. This could allow malicious content within the analyzed files to influence the agent's summary.
  • Ingestion points: Code content and git diffs are read from the filesystem and git history in scripts/risk-analyze.js (e.g., via collectDiff and expandUntrackedDir).
  • Boundary markers: The skill does not define explicit delimiters or instructions to prevent the agent from obeying commands embedded in the analyzed code.
  • Capability inventory: The skill possesses capabilities to read files and run git/grep commands, but lacks network access or filesystem write permissions.
  • Sanitization: No sanitization or filtering of code content is performed before the agent processes it for interpretation.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 24, 2026, 01:45 PM
Security Audit — agent-trust-hub — risk-assess