validate-and-fix

Pass

Audited by Gen Agent Trust Hub on May 16, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes local shell commands to manage project dependencies and run tests. Evidence includes calls to make install-deps, make test-unit, make test-integration, and make verify in SKILL.md. It also uses uv run black . and uv run ruff check . --fix to format code and fix linting issues.\n- [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface as it processes untrusted test output to determine subsequent actions.\n
  • Ingestion points: Test and linter output from make commands in SKILL.md.\n
  • Boundary markers: Absent.\n
  • Capability inventory: Shell command execution and file modification via make and uv tools in SKILL.md.\n
  • Sanitization: Absent.
Audit Metadata
Risk Level
SAFE
Analyzed
May 16, 2026, 11:32 AM
Security Audit — agent-trust-hub — validate-and-fix