The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill establishes a workflow where the agent ingests terminal output to influence its next actions, creating an indirect prompt injection surface. (1) Ingestion points: herdr pane read and herdr agent read in CLI-REFERENCE.md. (2) Boundary markers: The instructions lack requirements for explicit delimiters or instructions to ignore embedded commands in ingested output. (3) Capability inventory: The agent has access to herdr pane run and herdr agent prompt for executing commands and instructing other agents. (4) Sanitization: No data filtering or sanitization is specified.
[COMMAND_EXECUTION]: The capture-cli-help.py script uses subprocess.run to execute herdr CLI commands for help text discovery.
[EXTERNAL_DOWNLOADS]: The MAINTENANCE.md file provides instructions to fetch documentation and metadata from the herdrdev/herdr repository using the GitHub CLI.

herdr