The Agent Skills Directory

[COMMAND_EXECUTION]: The CLI logic in rhdh/cli.py and rhdh/config.py uses subprocess.run to interact with system tools such as git, gh, podman, docker, and acli for environment diagnostics and repository management.
[DYNAMIC_EXECUTION]: The rhdh local command in rhdh/cli.py resolves the path to an external rhdh-local executable at runtime and executes it, delegating arguments to this standalone tool.
[INDIRECT_PROMPT_INJECTION]: The skill manages and displays content from TODO.md and worklog.jsonl files. If these files are populated with malicious instructions from external sources (e.g., via a git commit or automated tool), they could influence the agent when displayed.
Ingestion points: The skill reads TODO.md (rhdh/todo.py) and worklog.jsonl (rhdh/worklog.py).
Boundary markers: None detected for these markdown/JSONL files.
Capability inventory: The skill can execute shell commands via subprocess.run for git and other tools.
Sanitization: The raw content of todos is printed to the agent context without specific sanitization of embedded instructions.

rhdh-context