cloudflare-workers-deployments
Pass
Audited by Gen Agent Trust Hub on Oct 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill incorporates high-quality security operating rules, including mandatory redaction of API tokens and secrets from all outputs, the use of
--dry-runto inspect JSON payloads before transmission, and a requirement for explicit user confirmation before performing any write operations (POST, PATCH, PUT, DELETE) on Cloudflare resources. - [INDIRECT_PROMPT_INJECTION]: The skill provides the agent with capabilities to read local project configuration files such as
package.jsonandwrangler.tomlto derive deployment commands. While this introduces an ingestion surface for untrusted data, the skill mitigates the risk by instructing the agent to verify scripts explicitly and requiring human approval before executing any derived commands. - Ingestion points: Local project files including
package.json, lockfiles, andwrangler.tomlare parsed to determine build scripts and project settings. - Boundary markers: None explicitly defined for file parsing, but the workflow mandates manual verification of derived commands.
- Capability inventory: Use of
cfandwranglerCLI tools, as well ascurlfor direct REST API interaction with Cloudflare services. - Sanitization: Mandatory use of
--dry-runand explicit user confirmation for all deployment-related tasks.
Audit Metadata