auto-board-task
Pass
Audited by Gen Agent Trust Hub on Jun 19, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: Indirect Prompt Injection analysis. The skill defines an indirect injection surface because it processes untrusted user data (GitHub project URLs and owner names) to be passed to downstream tools. However, the implementation is secure as it mandates the extraction of specific fields rather than forwarding raw natural language.
- Ingestion points: User-provided project coordinates (URL, owner, and number) described in Step 1 of SKILL.md.
- Boundary markers: The skill instructs the agent to format arguments as explicit
key=valuepairs, providing a structural boundary for theSkilltool invocation. - Capability inventory: The skill invokes the
workflowskill, which orchestrates sub-skills (gh-project-sync,task-agent) possessing file system, network (GitHub API), and code execution capabilities. - Sanitization: Instructions require the agent to resolve and extract project metadata rather than interpolating raw user messages directly into the workflow arguments.
- [SAFE]: Command Execution analysis. The skill uses the
Bashtool to resolve the absolute path for its bundled workflow YAML file. This is a limited and functional use of the shell for environment discovery (expanding<CLAUDE_PLUGIN_ROOT>) and does not involve untrusted input or elevated privileges.
Audit Metadata