security-express
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The audit scanner (
scripts/scan.sh) reads and displays raw lines from untrusted source code files to the agent. This creates a potential vector where malicious instructions embedded in comments or strings within the audited code could be interpreted by the agent during its review.\n - Ingestion points: The scanner script processes content from user-specified directories and files (e.g., package.json, .js, .ts files) via ripgrep.\n
- Boundary markers: The report generated by the script does not use explicit isolation markers to separate scanner output from the untrusted content being matched.\n
- Capability inventory: The skill utilizes shell commands (bash), file search (ripgrep), and regex matching to facilitate its security audit functions.\n
- Sanitization: No sanitization is applied to the extracted code snippets before they are presented to the agent.
Audit Metadata