squad-refine

Pass

Audited by Gen Agent Trust Hub on Jun 13, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructions include the use of curl for interacting with a task management API to retrieve (GET) and update (PATCH, POST) requirements. It also uses a bash subshell to execute a read_model helper script for internal configuration management.
  • [PROMPT_INJECTION]: The skill processes external data (task descriptions and implementation notes) which constitutes an indirect prompt injection surface. This is a low-risk exposure inherent to the skill's purpose of requirement analysis.
  • Ingestion points: Data is ingested from the /api/task/ endpoint, specifically the task title, description, tags, and implementation_notes (SKILL.md, Step ① and ① ½).
  • Boundary markers: The skill uses structured markdown templates for synthesis but does not employ specific delimiters or 'ignore' instructions for the data being processed.
  • Capability inventory: The skill can perform network operations via curl to modify task data and execute local model-resolution scripts.
  • Sanitization: There is no explicit sanitization described for the data retrieved from the API before it is processed by the model.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 13, 2026, 01:38 PM
Security Audit — agent-trust-hub — squad-refine