project-task-manager
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: Comprehensive analysis detected no malicious code, unauthorized command execution, or credential exfiltration attempts. The skill exclusively facilitates remote tool calls to the author's official domains (agentpmt.com) and uses standard installation procedures for the platform.
- [INDIRECT_PROMPT_INJECTION]: The skill accepts natural language objectives that are processed by an AI, which is an inherent surface for indirect prompt injection. 1. Ingestion points: 'objective' parameter in the generate action and 'task' parameter in the decompose action (SKILL.md). 2. Boundary markers: Absent in the current instruction set. 3. Capability inventory: Remote API calls for task tree management (generate, update, decompose, status, list). 4. Sanitization: Relies on the security and filtering mechanisms of the remote AgentPMT service.
Audit Metadata