kernel-write-dossier

Pass

Audited by Gen Agent Trust Hub on Jul 13, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No malicious patterns, obfuscation, or unauthorized data access attempts were detected. The skill focuses on orchestrating text generation and review workflows for business planning.\n- [PROMPT_INJECTION]: The skill processes untrusted external data (business ideas and note paths), representing a surface for indirect prompt injection. This is inherent to the skill's primary function and managed within standard execution boundaries.\n
  • Ingestion points: Source business idea and note paths (SKILL.md, section-generator.md)\n
  • Boundary markers: None present to isolate instructions from external data.\n
  • Capability inventory: File writing (SKILL.md) and skill orchestration (orchestrator.md).\n
  • Sanitization: None specified for input content.\n- [COMMAND_EXECUTION]: The skill instructs the agent to write finalized dossiers to user-provided output paths. This is a standard functional requirement for document assembly.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 13, 2026, 06:28 AM
Security Audit — agent-trust-hub — kernel-write-dossier