kb-init

Pass

Audited by Gen Agent Trust Hub on Aug 13, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill analyzes local workspace files such as the README, source code, and user notes to infer the domain and entities for a new knowledge bundle. This ingestion of untrusted project data creates a surface for indirect prompt injection.
  • Ingestion points: Project workspace files including README, existing documentation, code, and user notes as specified in step 2.
  • Boundary markers: The instructions do not define delimiters or specific warnings to ignore embedded instructions within the analyzed files.
  • Capability inventory: File system write access to create the directory structure and initialize multiple markdown files (index.md, log.md, spec/types.md, etc.) as detailed in step 3.
  • Sanitization: No explicit sanitization or validation of the inferred domain content is performed before it is written to the new documentation files.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 13, 2026, 08:35 AM
Security Audit — agent-trust-hub — kb-init