run-tests
Pass
Audited by Gen Agent Trust Hub on Aug 4, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill resolves and executes shell commands derived from the local repository's environment and configuration files such as Makefile, package.json, and various lockfiles.
- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by treating instructions found in the codebase as authoritative for command execution.
- Ingestion points: Reads commands and configuration from AGENTS.md, CLAUDE.md, .cursor/rules/*.mdc, CONTRIBUTING.md, Makefile, package.json, pyproject.toml, and various lockfiles.
- Boundary markers: Absent; the skill is instructed to trust and execute documented commands verbatim, providing a direct path from untrusted file content to shell execution.
- Capability inventory: Capable of executing arbitrary shell commands through various package managers and runners including uv, pnpm, bun, npm, make, and docker compose.
- Sanitization: No sanitization or validation logic is present to verify the safety of commands extracted from project documentation or scripts before execution.
Audit Metadata