engagement-report-council-critique
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- Processing of Untrusted Content: The skill is designed to ingest and analyze external data such as draft reports and research documents. While this represents a surface for indirect prompt injection, the skill includes explicit instructions to mitigate this risk.
- Ingestion points: The skill accepts several external inputs (draft, research, coverage, audience, and template) as specified in the Inputs section of SKILL.md.
- Boundary markers: The skill contains clear directives to treat these inputs as untrusted data and specifically warns against following any embedded directives within the supplied values.
- Capability inventory: The skill has the capability to write output to the local file system using the synthesis/ directory within a .working/ path.
- Sanitization: Robust path validation is required, including the rejection of parent traversal and absolute paths, to ensure the agent remains within a confined directory.
Audit Metadata