grants-management-manual
Pass
Audited by Gen Agent Trust Hub on Sep 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted external data, including award agreements, donor rules, budgets, and reporting calendars, as specified in the 'Required Inputs' table and workflow steps in
SKILL.md. This ingestion creates a potential attack surface for indirect prompt injection if malicious instructions are embedded within these documents. - Ingestion points: External donor agreements, budgets, and compliance rules are loaded into the agent context in
SKILL.md. - Boundary markers: There are no explicit delimiters or instructions provided to the agent to ignore or isolate instructions potentially contained within the ingested data.
- Capability inventory: The skill has the capability to generate document files (DOCX/PDF) and execute quality gate scans as detailed in
SKILL.md. - Sanitization: The workflow includes logic conformance scanning but lacks explicit text sanitization or filtering for the external inputs.
Audit Metadata