k2f-policy-document
Pass
Audited by Gen Agent Trust Hub on Oct 1, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill documentation includes a command to install a necessary dependency via
npx skills add suzheng/k2f. This resource is authored by the same vendor ('suzheng') and is integral to the skill's function. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and reformat user-supplied policy outlines and legacy text, which creates an attack surface for instructions embedded in that source data.
- Ingestion points: User source material, outlines, and legacy policy text.
- Boundary markers: The workflow mandates the use of 'Source' and 'Honesty' gates and requires labeling drafted sections with 'FOR HR/LEGAL REVIEW' to distinguish them from user-provided mandates.
- Capability inventory: The skill produces structured policy packages and supports exporting them to formats like PDF or DOCX using the referenced formatting engine.
- Sanitization: Instructions strictly forbid the invention of policy limits, legal citations, or jurisdictions, requiring verbatim use of user-supplied mandatory language.
Audit Metadata