The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill identifies and uses a local credential storage file at ~/.clawlink/credentials.json. While this is the tool's native mechanism, storing sensitive access keys in the home directory is a known exposure point.
[EXTERNAL_DOWNLOADS]: The skill instructs the agent to use npx to download and execute the @useclawlink/cli package from the NPM registry. This involves downloading and running code from a remote repository at runtime.
[COMMAND_EXECUTION]: The skill's primary functionality is delivered through shell commands (npx @useclawlink/cli ...), granting the agent the capability to execute external binaries and interact with the system shell.
[INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect injection because it interpolates untrusted user input directly into a shell command template.
Ingestion points: User natural language requests are passed into the <what you want to do> argument of the npx @useclawlink/cli actions discord command in SKILL.md.
Boundary markers: There are no explicit boundary markers or instructions provided to the agent to sanitize or escape the user input before command execution.
Capability inventory: The skill possesses the capability to execute shell commands (npx), access the network (via the CLI's Discord integration), and read/write to the local file system (credential management).
Sanitization: No sanitization or validation logic is defined to prevent command injection or parameter manipulation via the interpolated user input.

discord