inbound-plan-work

Pass

Audited by Gen Agent Trust Hub on Aug 4, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill documentation provides instructions for setting up the paperclipai CLI and the @bbengamin/paperclip-mcp-server via npm and npx. These resources are associated with the skill's author and represent standard components of the Paperclip environment.\n- [COMMAND_EXECUTION]: The skill utilizes the paperclipai CLI to manage Paperclip entities like issues, goals, and projects. These commands are administrative in nature and are only executed after the operator approves a proposed breakdown.\n- [PROMPT_INJECTION]: The skill processes data from external wiki pages and issue comments to inform its planning activities, creating a surface for indirect prompt injection. This is mitigated by a strict human-in-the-loop requirement for all mutations.\n
  • Ingestion points: Content fetched via paperclip-wiki-fetch, issue comments, and growth strategy artifacts.\n
  • Boundary markers: The process includes a mandatory approval rule ("Never create child issues until the operator approves the proposed breakdown") which prevents automated obedience to instructions embedded in data.\n
  • Capability inventory: Administrative operations including issue creation, goal updates, and wiki page management via confirmed plugin routes.\n
  • Sanitization: Instructions specify the removal of identifying details before public posting but do not explicitly detail sanitization or delimiting of the ingested content for prompt isolation.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 4, 2026, 08:26 PM
Security Audit — agent-trust-hub — inbound-plan-work