本文へ移動
cccskills
無料GitHub で公開

ai-workflow

Plan, verify, implement, and deliver repository work through evidence-backed ownership and quality gates.

インストール方法を見る

含まれるファイル(5)

  • SKILL.md2.3 KB
  • agents/openai.yaml265 B
  • references/evidence-and-verification.md500 B
  • references/prompt-patterns.md476 B
  • references/task-decomposition.md521 B

SKILL.md(原文)

インストールする前に、エージェントに与えられる指示の中身を確認できます。

AI Workflow

Establish the contract

Resolve the requested outcome, exclusions, allowed paths, and required checks. Verify repository owner, upstream, branch, and edit-marker scheme when the work changes inherited files, ownership, project references, synchronization, or automation. Track commit count and remote branch SHA only when the task includes committing, pushing, or publishing a PR.

Record the initial worktree state before editing. Inspect SPDX changes only when game code, resources, or configuration are in scope.

Complete the preflight

Fill in only fields relevant to the current task. Mark the others not applicable instead of investigating unrelated surfaces.

Repository and owner tag:
Owner module and underscore paths:
Inherited paths and marker syntax:
Target and caller assemblies:
Required declarations and access:
Existing extension point:
Existing test owner:
Localization owner and requested locale files:
Verification plan:

No implementation starts while a correctness-critical field for the current task is unknown. Skip fields that do not apply rather than performing unrelated discovery.

Explore before designing

Map entry points, dependencies, resources, tests, lifecycle, localization order, and current behavior using exact paths and symbols.

Subagent summaries are not proof without declarations, file evidence, or command output.

Implement narrowly

Reuse existing owners and infrastructure. Mark inherited edits with the current repository marker. Do not mark owner-local module or underscore paths. Do not edit SPDX.

Treat English localization as the structural source of truth. Update Russian only when explicitly requested; do not mirror English structural changes automatically.

Verify claims

Run the narrowest meaningful check first, then broaden according to risk. A missing tool is a limitation, not a pass.

Deliver

Inspect final status and diff. Confirm the applicable scope, marker placement, accessibility, requested localization coverage, duplicate infrastructure, and exact verification results. Confirm commit count and remote branch SHA only if commits or remote publication were part of the task.

レビュー

まだレビューはありません。使ってみた感想をお寄せください。

同じリポジトリのスキル

概要と使いどころ

Implement actions, cooldowns, DoAfter flows, interruption, cancellation, and authoritative completion safely.

日本語の概要は準備中です。原文の説明を表示しています。

ArcaneSS14/arcane-new32026年10月12日 更新

Implement privileged commands, admin UI, role checks, audit logs, and safe operational controls.

日本語の概要は準備中です。原文の説明を表示しています。

ArcaneSS14/arcane-new32026年10月12日 更新

Implement replicated appearance data, GenericVisualizer mappings, sprite layers, overlays, and client visual systems.

日本語の概要は準備中です。原文の説明を表示しています。

ArcaneSS14/arcane-new32026年10月12日 更新

atmos

無料

Modify atmos simulation, gases, fire, pressure, pipes, devices, UI, and hot paths safely.

日本語の概要は準備中です。原文の説明を表示しています。

ArcaneSS14/arcane-new32026年10月12日 更新

audio

無料

Add data-driven audio with correct prediction, audience, resources, attribution, and verification.

日本語の概要は準備中です。原文の説明を表示しています。

ArcaneSS14/arcane-new32026年10月12日 更新

Implement BUI contracts, server validation, localized client windows, lifecycle cleanup, and state refresh.

日本語の概要は準備中です。原文の説明を表示しています。

ArcaneSS14/arcane-new32026年10月12日 更新

ArcaneSS14 のスキルをすべて見る

このスキルの問題を報告する