The Agent Skills Directory

[COMMAND_EXECUTION]: The CLI tool utilizes subprocess.run to execute various local system commands for environment diagnostics. These include git for branch and status checks, gh for GitHub CLI authentication and API access, and container runtimes like podman or docker. It also supports a compatibility seam that executes a standalone rhdh-local binary resolved from the environment or search paths.
[INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection by processing external project files. * Ingestion points: Reads repository metadata from backstage.json and source.json, and processes user-generated content from TODO.md and worklog.jsonl. * Boundary markers: Absent; files are read and parsed without explicit delimiters or instructions to ignore embedded commands. * Capability inventory: The skill has broad command execution capabilities via subprocess.run for system tools and the rhdh-local delegate. * Sanitization: Content is parsed as JSON or via regex for specific fields, but the resulting values are not sanitized for potential natural language instructions embedded within the files.

rhdh-context