The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a local Python script located at .claude/skills/gh/scripts/github_project_setup.py using the uv run command to perform bulk label updates and Project V2 status changes.
[COMMAND_EXECUTION]: The milestone number is passed as a command-line argument to the script. The skill workflow effectively mitigates command injection risks by first resolving the number against the official GitHub API via MCP tools before passing it to the shell environment.
[INDIRECT_PROMPT_INJECTION]: The skill processes external content (milestone titles, issue titles, and labels) fetched from the GitHub API.
Ingestion points: External data enters the context via mcp__plugin_dh_backlog__backlog_list_milestones and mcp__plugin_dh_backlog__backlog_list_issues calls described in SKILL.md.
Boundary markers: The skill implements a manual confirmation step (AskUserQuestion) in Step 3, ensuring the user reviews the specific issues to be transitioned before any scripts are executed.
Capability inventory: The skill possesses the capability to execute shell commands (uv run) and perform write operations to GitHub via the provided automation script.
Sanitization: The skill avoids direct interpolation of raw user input into shell commands by requiring a resolution step where the input must match an existing milestone ID returned by the GitHub API.

start-milestone