decision-mapping

Pass

Audited by Gen Agent Trust Hub on Jun 18, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by design.
  • Ingestion points: It reads an external markdown file (the 'decision map') from the project directory during the 'Resume' flow.
  • Boundary markers: The instructions explicitly state to 'Load the whole map as context' without defining delimiters or instructions to ignore potential commands embedded within that content.
  • Capability inventory: Based on the context loaded from the map, the skill is authorized to invoke further skills such as /prototype (which generates/executes code), /grilling, and /domain-modelling.
  • Sanitization: There is no mention of sanitization, validation, or escaping of the content retrieved from the markdown file before it is processed by the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 18, 2026, 08:31 AM
Security Audit — agent-trust-hub — decision-mapping