codebase-context
Pass
Audited by Gen Agent Trust Hub on Mar 10, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [SAFE]: The skill utilizes a restricted set of read-only shell commands (ls, cat, find, grep, wc) to inspect project organization and metadata, which is appropriate for its stated objective of codebase analysis.\n- [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface as it ingests content from external files within the analyzed codebase.\n
- Ingestion points: External project files like package.json, requirements.txt, and README.md are read using cat and head commands.\n
- Boundary markers: No explicit delimiters or instructions to ignore instructions within the read content are present in the skill definition.\n
- Capability inventory: The skill possesses the ability to execute find, head, cat, ls, and wc via Bash, and uses the Grep and Read tools.\n
- Sanitization: No sanitization or validation of the content of the read files is performed before the data is passed to the agent context.
Audit Metadata