agent-goal-planner

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill architecture is designed to ingest and act upon complex user-defined goals to orchestrate automated tasks. \n
  • Ingestion points: The task field in tool orchestrations and objectives provided to the GOAP planner in SKILL.md. \n
  • Boundary markers: The instructions do not specify delimitation or isolation protocols to prevent embedded instructions in processed tasks from overriding the planner's logic. \n
  • Capability inventory: The skill is configured to use orchestration, swarm initialization, and memory storage tools (mcp__claude-flow__task_orchestrate, mcp__claude-flow__swarm_init, mcp__claude-flow__memory_usage). \n
  • Sanitization: There are no explicit instructions for validation or sanitization of input data before it influences the planning and tool-calling workflow. \n- [NO_CODE]: The skill consists solely of descriptive instructions and metadata in SKILL.md without accompanying scripts or executables.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 11:38 AM
Security Audit — agent-trust-hub — agent-goal-planner