メニュー
Planning-centric workflow for substantial work across software, analyses, artifacts, documentation, and configuration. Invoke manually with /work-cycle when alignment matters; not for straightforward one- or two-edit requests.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Planning-centric workflow for substantial work across software, analyses, artifacts, documentation, and configuration. Invoke manually with $work-cycle when alignment matters; not for straightforward one- or two-edit requests.
Create or update developer-facing and agent-facing project documentation, or run module-scoped codebase audits. Invoke manually with $documentation for codebase maps, module documentation, codebase review/audit workflows, and README.md or AGENTS.md work.
Documentation subskill for project hygiene maintenance. Use through $documentation when the user asks to maintain, clean up, health-check, refresh AGENTS.md/notebook/TODOs, inspect git branches/PRs/stashes, or review overall project state. Do not use for codebase audit, codebase review, release-check, or module-scoped bug-finding requests; those belong to module-codebase-review.
Manual-only workflow for codebase audit, codebase review, release-check, paper-freeze-check, and module-scoped multi-agent read-only bug/code-smell audits that produce a prioritized notebook report and TODO-ready findings. Use this instead of maintain-project for requests that ask to audit or review code behavior.
Ergonomic notebook-backed TODO capture for deferred work. Invoke manually with /defer when the user asks to defer, save, park, or add a TODO for later work while preserving the current conversation context.
Initialize a project workspace for Claude Code with project instructions, repository checks, notebook scaffolding, optional remotes, and optional O2/SLURM setup. Invoke manually with /init-project for first-time project setup, notebook bootstrap, project remotes, or O2/SLURM setup.
| name | work-cycle |
| description | Planning-centric workflow for substantial work across software, analyses, artifacts, documentation, and configuration. Invoke manually with /work-cycle when alignment matters; not for straightforward one- or two-edit requests. |
| recommended_scope | global |
| disable-model-invocation | true |
Invoke this skill manually with /work-cycle. Do not rely on automatic trigger behavior.
This skill coordinates substantial work through a question-plan-implement-review cycle. It is domain-neutral at the top level; load the relevant domain reference before planning:
references/domains/software.md.references/domains/analysis.md./artifacts and its relevant subskill.The user may request one or more special modes, in which case read and follow the corresponding subprompt:
references/modes/worktree.mdreferences/modes/afk.mdreferences/modes/methods-first.mdreferences/modes/grill-me.mdSTOP: if you are not in Plan Mode and the user did not explicitly say AFK, immediately ask the user to switch you into Plan Mode with Shift+Tab, then wait.
Most substantial work should go through a planning phase in which you go back and forth and align with the user. Planning must occur in Plan Mode unless AFK mode is explicit.
A shorthand request such as todo 22, an issue number, or a named skill is not by itself implementation approval and is not a substitute for Plan Mode.
During planning, loop at least once through:
Draft and edit plans in notebook/plans/. The plan should include:
After drafting the plan, assess whether a different approach would be simpler or better aligned with user intent; if so, ask an additional question and iterate.
Be open to issues with the plan or better approaches during implementation. If the plan is deeply flawed, stop and ask for user input; otherwise, use your best judgment and report afterward what changed from the plan and why.
Run relevant validation and iteratively debug. If the task is substantial, or if you are in worktree mode, trigger external review when complete and report findings to the user.
Create or update notebook entries for new analyses, code changes that alter interpretation of scientific results, codebase investigations, significant decisions, or substantial work whose context should survive the session. For entry creation after implementation, fork a subagent.
If the user asked to work on a specific notebook TODO, complete the TODO once implementation is finished in normal mode, or as soon as the PR is opened in worktree mode; do not wait for merge.