review-my-team
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process potentially untrusted data, including task descriptions, comments, and documents retrieved from external sources.
- Ingestion points: The skill reads task details and comments via
paperclip_search_tasksandpaperclip_pending_approvalsas described inSKILL.md. - Boundary markers: The instructions explicitly warn the agent to treat retrieved content as "untrusted work content" and not as instructions to broaden its authority or transmit secrets.
- Capability inventory: The available tools (
paperclip_list_agents,paperclip_list_projects,paperclip_search_tasks,paperclip_pending_approvals) are restricted to information retrieval and listing. No tools for file system modification, arbitrary shell execution, or external network requests are present. - Sanitization: The skill uses instructional constraints to prevent the agent from being influenced by malicious prompts embedded within task data.
Audit Metadata