paperclip-plan-work
Pass
Audited by Gen Agent Trust Hub on Aug 4, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: Executes shell commands through the
paperclipaiCLI andcurlfor the purpose of managing Paperclip goals, projects, and issues. - [EXTERNAL_DOWNLOADS]: References the
paperclipaiglobal package and executes the@bbengamin/paperclip-mcp-serverpackage vianpxfor fallback API operations. - [DATA_EXFILTRATION]: Communicates with an external Paperclip API endpoint to synchronize and update organizational data.
- [PROMPT_INJECTION]: Processes data from parent issues, plan documents, and wiki pages which could contain untrusted instructions.
- Ingestion points: Reads issue bodies, comments, and external wiki source material via
paperclip-wiki-fetch. - Boundary markers: Proposes breakdowns in a structured format and requires explicit operator approval before performing any mutations.
- Capability inventory: Has the ability to create and modify issues, projects, and goals via the Paperclip control plane.
- Sanitization: Relies on manual human review of proposed breakdown structures as the primary safeguard.
Audit Metadata