cloudflare-deploy
Pass
Audited by Gen Agent Trust Hub on Sep 10, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes and deploys user-provided code and configurations using the Wrangler CLI and Cloudflare APIs (SKILL.md). • Ingestion points: Deployment requests involving user-supplied source code and environment variables (SKILL.md). • Boundary markers: None identified in the skill's instructional logic to isolate user code from the agent's deployment routine. • Capability inventory: Supports full deployment capabilities, including command execution (wrangler deploy) and file system writes (SKILL.md, references/wrangler/README.md). • Sanitization: Reference documentation provides guidance on preventing SQL injection in D1, but the skill lacks mechanisms to sanitize code before deployment.
- [EXTERNAL_DOWNLOADS]: The skill documentation includes patterns for downloading environment setup scripts from well-known vendor domains, such as Coder (references/sandbox/patterns.md). It also references numerous official libraries from npm and PyPI for platform integration.
- [COMMAND_EXECUTION]: The skill uses the Wrangler CLI for application deployment, secret management, and resource provisioning (SKILL.md). It includes patterns for executing system commands within isolated Cloudflare sandboxes (references/sandbox/api.md).
Audit Metadata