The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses spawnSync to execute the system curl binary. This is used for multipart/form-data uploads of Worker scripts and binary data transfers for R2 storage, which are standard requirements for interacting with these specific Cloudflare API endpoints.
[INDIRECT_PROMPT_INJECTION]: The skill acts as a proxy for deploying code and data to a cloud environment, which creates an attack surface for indirect injection if the source data is compromised.
Ingestion points: Local files are read in scripts/workers.ts (scripts), scripts/r2-storage.ts (binary objects), and scripts/kv-storage.ts (JSON data).
Boundary markers: The skill does not use specific delimiters or protective instructions for the content of the files it processes.
Capability inventory: Significant capabilities include deploying code to Workers, modifying DNS records, and writing to cloud storage.
Sanitization: The skill validates resource names (Workers, buckets, namespaces) using regex and encodes URL path segments using encodeURIComponent to prevent path traversal or API manipulation.
[SAFE]: Several security best practices are implemented to protect the user's environment:
Temporary files containing worker code and sensitive API metadata are created with restrictive file permissions (0o600), ensuring they are not accessible by other users on the system.
The validation cache is also stored with restrictive permissions (0o600).
The skill uses token-based authentication via environment variables and explicitly warns against committing secrets to version control.
Destructive operations, such as deleting buckets or namespaces, require manual confirmation from the user unless an explicit override flag is provided.

cloudflare-manager