kaizen-improvement-system

Pass

Audited by Gen Agent Trust Hub on Sep 18, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external, untrusted documents such as RFPs, evaluation criteria, and proposal drafts, which could contain malicious instructions.
  • Ingestion points: The 'Required Inputs' table in SKILL.md identifies ToR/RFP and proposal artefacts as primary data sources.
  • Boundary markers: The skill does not explicitly define text delimiters or 'ignore' instructions for external content, although it workflow includes a 'quarantine' step for unsupported claims.
  • Capability inventory: The skill utilizes file system 'read' and 'search' capabilities and interacts with other skill modules like 'digital-research-skills'.
  • Sanitization: The workflow includes 'anti-slop' audits, red-team simulations, and 'evaluator journey' checks to ensure that generated content is verified against evidence and compliance rules.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 18, 2026, 03:41 AM
Security Audit — agent-trust-hub — kaizen-improvement-system