paperclip-admin

Pass

Audited by Gen Agent Trust Hub on Aug 4, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill installs the paperclipai CLI and executes the @bbengamin/paperclip-mcp-server package from the NPM registry. These resources are provided by the skill author for legitimate administrative functions.
  • [COMMAND_EXECUTION]: The skill uses the paperclipai CLI and curl to perform control-plane operations and interact with the Paperclip REST API.
  • [DATA_EXFILTRATION]: Network operations are limited to communicating with the Paperclip API endpoint (defaulting to localhost:3100) or vendor documentation sites, following user-configured settings.
  • [PROMPT_INJECTION]: The skill ingests external data from the Paperclip API, which serves as an indirect prompt injection surface. 1. Ingestion points: Data is read via paperclipai CLI output and API responses. 2. Boundary markers: No specific prompt delimiters are used for ingested data, but the skill mandates a human-in-the-loop approval process for all state-changing actions. 3. Capability inventory: The skill can execute shell commands and modify agent instructions. 4. Sanitization: Relies on manual operator verification and explicit approval steps.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 4, 2026, 08:26 PM
Security Audit — agent-trust-hub — paperclip-admin