The Agent Skills Directory

[PERSISTENCE]: The skill contains instructions to modify the user's shell profile file to persist sensitive environment variables across sessions.
Evidence: The 'Optional persistent setup' section in SKILL.md provides a command to append the Cloudflare API token to ~/.zshrc using echo "export CLOUDFLARE_API_TOKEN='YOUR_TOKEN'" >> ~/.zshrc.
[CREDENTIALS_UNSAFE]: The skill encourages the storage of sensitive API tokens in plain text within shell configuration files, which is a less secure practice compared to utilizing dedicated secret management tools or encrypted environments.
Evidence: Instructions for persistent setup in SKILL.md directly export secrets into a hidden shell configuration file.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from the user and interpolates it into shell commands and API requests, creating a surface for indirect injection if malicious values are provided.
Ingestion points: The skill processes user-supplied values for domain/host, record type, and target value in SKILL.md.
Boundary markers: No delimiters or specific instructions to ignore embedded commands within user-provided strings are present.
Capability inventory: The skill utilizes urllib in Python for network operations, shell commands like dig and curl, and file-writing capabilities via shell redirection.
Sanitization: There is no evidence of input validation, escaping, or sanitization for the variables used in the Python script or shell verification commands.

cloudflare-dns