parsing-source
Pass
Audited by Gen Agent Trust Hub on Aug 22, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The documentation in SKILL.md states that language parser libraries are downloaded on first use and cached locally. This is a functional requirement for the ts-pack tool to support various languages.
- [PROMPT_INJECTION]: The skill enables the agent to ingest and parse source code from files or stdin, which represents an indirect prompt injection attack surface where embedded instructions could influence the agent. 1. Ingestion points: Source files and stdin input provided to the
ts-pack parsecommand. 2. Boundary markers: Not present; the instructions do not include delimiters or warnings to ignore embedded content. 3. Capability inventory: The skill facilitates file reading and the generation of syntax tree data. 4. Sanitization: The tool provides a structural analysis of the input but does not filter or sanitize the content within the code nodes.
Audit Metadata