ast-code-graph
Pass
Audited by Gen Agent Trust Hub on Mar 30, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill invokes the ast-grep (or sg) CLI to perform structural searches and rewrites within the user's codebase.
- [COMMAND_EXECUTION]: Executes bundled scripts scripts/build-graph.py (Python) and scripts/parse-js.mjs (Node.js) to construct code graphs and extract symbol information.
- [COMMAND_EXECUTION]: Accesses and reads local source code files to generate structural metadata; no network activity or exfiltration of the analyzed code was observed.
- [EXTERNAL_DOWNLOADS]: Recommends the installation of ast-grep and Node.js libraries (acorn, acorn-walk) from official registries (NPM, Homebrew, Cargo) to support its parsing capabilities.
- [PROMPT_INJECTION]: As the skill ingests and analyzes external source code, it presents an indirect prompt injection surface where malicious code comments could attempt to influence agent behavior. (Ingestion points: Local source code files; Boundary markers: None; Capability inventory: File system read and CLI tool execution; Sanitization: Structural parsing via AST without execution).
Audit Metadata