clickclack

Pass

Audited by Gen Agent Trust Hub on Sep 16, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands for deployment and infrastructure management, including git pull, pnpm deploy:cloudflare, and wrangler containers info.
  • [DATA_EXPOSURE]: The instructions contain infrastructure identifiers for the Cloudflare environment, specifically a Cloudflare Account ID (91b59577e757131d68d55a471fe32aca), a Zone ID (e7f93ba1ec24d75ce60c468bbaba5cb5), and a legacy server IP address (157.90.237.80). These are documented for operational purposes and do not include authentication secrets.
  • [INDIRECT_PROMPT_INJECTION]: The skill uses curl to fetch content from the production URLs (e.g., https://clickclack.chat/) to verify deployment hashes. This creates a surface for indirect prompt injection if the target URL serves malicious content intended to influence the agent.
  • Ingestion points: curl -s https://clickclack.chat/ (found in SKILL.md).
  • Boundary markers: None; the output is piped directly to grep.
  • Capability inventory: git, docker, pnpm, wrangler (found in SKILL.md).
  • Sanitization: None; the skill relies on exact regex matching via grep to extract script hashes.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 16, 2026, 11:26 PM
Security Audit — agent-trust-hub — clickclack