apply
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill has a surface for indirect prompt injection because it ingests and processes untrusted external data.
- Ingestion points: Job description (JD) text or files and external research links are accepted as input in the
argument-hintand processed throughout the pipeline inSKILL.md. - Boundary markers: The instructions do not define specific delimiters or "ignore instructions" wrappers for the external JD text when it is passed to sub-skills like
/matchor/tailor. - Capability inventory: The skill possesses significant local capabilities, including writing multiple file types (PDF, DOCX, HTML, YAML, MD) to the filesystem and performing web searches via the
/researchsub-skill. - Sanitization: No explicit sanitization, filtering, or validation of the input JD text or research content is described before it is interpolated into the workflow.
Audit Metadata