ln-510-test-planner
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- EXTERNAL_DOWNLOADS (LOW): The file
diagram.htmlloads the Mermaid.js library from an external CDN (https://cdn.jsdelivr.net/npm/mermaid@10/dist/mermaid.min.js). While Mermaid is a reputable tool, fetching scripts from external domains at runtime is a dependency risk. - PROMPT_INJECTION (LOW): The skill is vulnerable to indirect prompt injection (Category 8) as it parses untrusted external data to drive its orchestration logic.
- Ingestion points: The skill reads Linear comments to check for research and testing status in Phase 2 and Phase 3.
- Boundary markers: The skill relies on specific string headers (e.g., '## Test Research:') which can be spoofed by any user with comment access to the Linear story, potentially tricking the orchestrator into skipping mandatory testing phases.
- Capability inventory: The orchestrator has the ability to invoke multiple sub-agents using the
Tasktool. - Sanitization: The skill mitigates risk by using
subagent_type: "general-purpose"for worker invocation, which provides context isolation between the orchestrator and the specialized workers.
Audit Metadata