The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses curl to send HTTP POST requests for notifications and cat to read its configuration file at ~/.claude/feishu.json.
[EXTERNAL_DOWNLOADS]: Communicates with remote Feishu API endpoints and user-configured bridge servers to deliver messages and poll for responses.
[PROMPT_INJECTION]: An indirect prompt injection surface is present in the interactive mode workflow. 1. Ingestion points: Remote user input is retrieved via the poll endpoint of the $BRIDGE_URL specified in SKILL.md. 2. Boundary markers: No delimiters or instructions to ignore embedded commands are used when processing the polled data. 3. Capability inventory: The skill uses curl and cat and facilitates data transfer between remote endpoints and other agent skills. 4. Sanitization: The skill does not perform validation or sanitization on the data received from the external bridge before returning it to the agent context.

feishu-notify