work
Work Execution Command
Execute work efficiently while maintaining quality and finishing features.
Introduction
This command takes a work source — a feature issue carrying a plan comment (the <!-- tunan:plan --> marker comment on the feature issue #N) or a bare prompt describing the work — and executes it systematically. The focus is on shipping complete features by understanding requirements quickly, following existing patterns, and maintaining quality throughout.
A feature is one GitHub issue for its lifetime: the requirement is the issue body, and the plan lands as a comment on that same feature issue (marker <!-- tunan:plan -->, label tunan:plan). There are no separate plan or solution issues. The feature issue #N is the durable handle passed across the pipeline; on shipping, the hand-off to compound adds a tunan:solution comment to the same feature issue. Pass and link by feature issue NUMBER/URL and #<N> references — never read or write a plan/solution as a local file. Read references/comment-chain-storage.md for the model and the exact gh recipes.
Interaction Method
At any point where work asks the user to choose among options (e.g., the branch-rename decision in Phase 1, or any fork in how to proceed), use the platform's blocking question tool: AskUserQuestion in Claude Code (call ToolSearch with select:AskUserQuestion first if its schema isn't loaded), request_user_input in Codex, ask_user in Gemini, ask_user in Pi (requires the pi-ask-user extension). Do not improvise a numbered list in chat as the default — that is the fallback only when no blocking tool exists or the call errors. When more than 4 distinct options are relevant, still fire the tool: present the 4 most important and name the rest in the question stem for free-form selection — AskUserQuestion caps at 4, and option count never justifies a numbered list. Never silently skip the question. Load the align skill for the full protocol.
Input Document
The input document for this run is the input this skill was invoked with — present in the current prompt or conversation, whether the user provided it directly or a calling skill passed it (e.g. lfg in mode:pipeline, which passes a plan path). It may be a plan or spec path, a mode: token followed by a path, or a bare work prompt. The rest of this skill refers to it as <input_document>; if nothing was provided, treat <input_document> as blank.