correct-course

Pass

Audited by Gen Agent Trust Hub on Jul 27, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill ingests data from local project files, such as tasks.md, docs/product/vision.md, and .skills/corrections.md, to diagnose plan deviations. These files could potentially contain embedded instructions that aim to influence the agent's behavior.
  • Ingestion points: The skill reads tasks.md, docs/product/vision.md, and .skills/corrections.md (SKILL.md).
  • Boundary markers: There are no explicit text delimiters mentioned for the ingested file content, but the skill enforces two separate human-in-the-loop approval gates (Phase 1 Diagnosis and Phase 2 Change Proposal) before proceeding.
  • Capability inventory: The skill can invoke other sub-skills (write-plan, write-design, write-requirements, sync-spec, domain-modeling) which have the capability to modify project documentation and plan files.
  • Sanitization: No specific sanitization or filtering of the ingested file content is mentioned.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 27, 2026, 02:10 PM
Security Audit — agent-trust-hub — correct-course