skills/clawlink-hq/skills/agencyzoom/Gen Agent Trust Hub

agencyzoom

Pass

Audited by Gen Agent Trust Hub on Jun 28, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSCREDENTIALS_UNSAFEPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses npx to execute commands for listing, describing, and running AgencyZoom actions via the ClawLink CLI.
  • [EXTERNAL_DOWNLOADS]: The @useclawlink/cli package is fetched from the public NPM registry, which is a well-known service for this vendor's tools.
  • [CREDENTIALS_UNSAFE]: The skill documents that authentication credentials are saved to ~/.clawlink/credentials.json. This is standard behavior for the CLI tool's local session management.
  • [PROMPT_INJECTION]: The skill has an indirect injection surface as it reads data from an external CRM. 1. Ingestion points: CRM data (leads, notes) enter the context through CLI command output (SKILL.md). 2. Boundary markers: Absent; no specific markers isolate retrieved data. 3. Capability inventory: The agent can execute shell commands via npx (SKILL.md). 4. Sanitization: The skill advises user confirmation before writes, which serves as a manual review gate.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 28, 2026, 02:42 PM
Security Audit — agent-trust-hub — agencyzoom