justfile
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted local files (like package.json, mix.exs, or fly.toml) to determine project settings and script names, which could be exploited to influence the generated justfile if these files contain malicious payloads.
- Ingestion points: Reading of project files including mix.exs, package.json, pyproject.toml, requirements.txt, go.mod, Cargo.toml, fly.toml, and Dockerfile (SKILL.md, Instructions 1 and 2).
- Boundary markers: None identified to isolate project file contents from the generation instructions.
- Capability inventory: The skill performs file-write operations to create the justfile in the project root.
- Sanitization: No explicit logic is provided to sanitize or validate strings extracted from project files before they are included in the output.
- [COMMAND_EXECUTION]: The skill generates shell commands intended for execution by the user through the 'just' utility.
- Evidence: SKILL.md provides templates for various ecosystems containing commands like 'npm run dev', 'mix phx.server', 'cargo run', and 'rm -rf' operations.
Audit Metadata