outreach-record-run

Pass

Audited by Gen Agent Trust Hub on Aug 4, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONCREDENTIALS_UNSAFEREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill documentation instructs users to install the paperclipai package globally via npm and provides configuration for the Claude CLI to add an MCP server.
  • [REMOTE_CODE_EXECUTION]: The skill documentation recommends running the @bbengamin/paperclip-mcp-server using npx -y, which downloads and executes code from the npm registry at runtime.
  • [COMMAND_EXECUTION]: The skill extensively uses shell commands via the paperclipai CLI and curl to manage resources, perform health checks, and interact with the Paperclip REST API.
  • [CREDENTIALS_UNSAFE]: The skill is designed to interact with sensitive authentication data stored in ~/.paperclip/auth.json and ~/.paperclip/context.json. It includes explicit warnings to avoid printing or exposing bearer tokens during execution and recommends using environment variables for secure token management.
  • [PROMPT_INJECTION]: The skill handles untrusted data from Paperclip wiki pages and user-provided specifications. It mitigates indirect prompt injection risks by enforcing a human-in-the-loop approval boundary for all mutations and implementing a mandatory verification step that reads back records after every write.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 4, 2026, 08:26 PM
Security Audit — agent-trust-hub — outreach-record-run