new-command
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from GitHub issues to drive a multi-phase code and documentation generation workflow. Because the agent is instructed to 'read the GitHub issue thoroughly' and use its content to implement command logic and unit tests, a malicious issue could contain embedded instructions designed to trick the agent into including malicious functionality or sensitive data exfiltration logic in the generated source code.
- Ingestion points: User-provided GitHub issue URLs and specification text (Phase 1).
- Boundary markers: Absent. The instructions do not provide delimiters or warnings to the agent to ignore instructions found within the processed issue content.
- Capability inventory: The skill possesses the capability to write new TypeScript and Markdown files to the local file system (Phase 2, 3, and 4) and execute shell commands via
npm testandnpm run build(Phase 3 and 5). - Sanitization: Absent. The skill lacks steps to validate or sanitize the external input before it is interpolated into the code generation process, relying on the agent to manually extract parameters without verifying intent.
Audit Metadata