council-forge / template
AGENTS.md0 starsChanged 6 months agoMITView on GitHub
本檔案是 artifact-first multi-agent workflow 的文件索引。每個 agent 只需載入自己的入口檔 + 當前階段所需的參考文件。
In this repository14 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- template/CLAUDE.mdCLAUDE.md
- .github/copilot-instructions.mdCopilot instructions
- acquire-codebase-knowledgeSkill
- agent-governanceSkill
- bootstrap-prompt-builderSkill
- code-tourSkill
- doublecheckSkill
- mcp-security-auditSkill
- quality-playbookSkill
- security-auditSkill
- security-reviewSkill
- suggest-awesome-github-copilot-skillsSkill
The file
Files
AGENTS -- 文件索引
本檔案是 artifact-first multi-agent workflow 的文件索引。每個 agent 只需載入自己的入口檔 + 當前階段所需的參考文件。
Agent 入口檔
| Agent | 入口檔 | 角色 | 自動載入 |
| Claude Code | CLAUDE.md | 協調者 | Yes (project instruction) |
| Gemini CLI | GEMINI.md | 研究與 read-only memory curator | Yes (passed via prompt) |
| Codex CLI | CODEX.md | 實作主責 | Yes (passed via prompt) |
文件模組
| File | 用途 | ~Tokens | 載入時機 |
docs/orchestration.md | 系統提示骨幹:目標、原則(§1-§2.8) | 1100 | Claude:session 開始 |
docs/orchestration-workflow.md | workflow/gate/PDCA 細節與 sync contract(§3+) | 1400 | Claude:流程與 gate 判定前 |
docs/artifact_schema.md | schema 入口與通用規則(§1-§4 + §5索引) | 900 | 寫任何 artifact 前 |
docs/schemas/artifact-spec-task.md | Task schema(原 §5.1) | 450 | 撰寫 task artifact 前 |
docs/schemas/artifact-spec-research.md | Research schema(原 §5.2) | 650 | 撰寫 research artifact 前 |
docs/schemas/artifact-spec-plan.md | Plan schema(原 §5.3) | 550 | 撰寫 plan artifact 前 |
docs/schemas/artifact-spec-code.md | Code schema(原 §5.4) | 900 | 撰寫 code artifact 前 |
docs/schemas/artifact-spec-test.md | Test schema(原 §5.5) | 350 | 撰寫 test artifact 前 |
docs/schemas/artifact-spec-verify.md | Verify schema(原 §5.6) | 800 | 撰寫 verify artifact 前 |
docs/schemas/artifact-spec-decision.md | Decision schema(原 §5.7) | 700 | 撰寫 decision artifact 前 |
docs/schemas/artifact-spec-status.md | Status schema(原 §5.8) | 500 | 更新 status artifact 前 |
docs/schemas/artifact-gallery.md | 跨類型範例集(源自 §5) | 250 | 需要快速對照範例時 |
docs/subagent_roles.md | 角色責任敘述(§1, §3-§9) | 2200 | 派發 subagent 前 |
docs/raci-matrix.md | RACI 與 capability 矩陣(原 §2) | 900 | 派發 subagent 前 |
docs/workflow_state_machine.md | 8 個狀態 + 合法轉移 | 600 | 狀態轉移前 |
docs/premortem_rules.md | 風險分析格式 + 品質護欄 | 1900 | 進入 coding gate 前 |
docs/red_team_runbook.md | 紅隊演練 runbook:靜態攻擊、live drill、復盤流程 | 1500 | 紅隊演練前 |
docs/red_team_scorecard.md | 紅隊演練評分矩陣與總結判定 | 900 | 演練記錄與復盤時 |
docs/red_team_backlog.md | 紅隊演練後續補強清單 | 700 | 復盤 / 補強規劃時 |
docs/subagent_task_templates.md | Template 索引(指向 docs/templates/) | 200 | 派發 subagent 時 |
docs/templates/<role>/TEMPLATE.md | 各角色 prompt 範本(含 YAML frontmatter) | 每個 ~150 | 派發對應 subagent 時 |
artifacts/scripts/discover_templates.py | Template auto-discovery CLI | 200 | 派發 subagent 前 |
docs/lightweight_mode_rules.md | 小任務精簡流程規則 | 350 | lightweight mode 任務時 |
Markdown 書寫語言規範
- 長期維護的 Markdown 文件以繁體中文(臺灣)為主。
- 專有名詞、檔名、CLI 指令、環境變數、
artifact type、狀態值、placeholder、schema literal 保留英文原字。 - 不得更動會被 agent、validator、腳本依賴的精確字串,例如
## Metadata、Task ID、Artifact Type、Owner、Status、Last Updated與各種狀態值。 - 所有規範中的紀錄時間、
Last Updated與相關時間戳,一律使用Asia/Taipei,採 ISO 8601 並帶+08:00。 - source template repo(含
.council-forge-source-repo)中,root、template/與 Obsidian 入口文件必須保持語義一致;由template/複製出的 downstream terminal repo 不再建立新的template/,只維護 root 文件與OBSIDIAN.md。 - GitHub 對外入口以
README.md/README.zh-TW.md為準;Obsidian 入口以OBSIDIAN.md為準。 - 歷史 artifacts、實驗輸出、外部 repo 內 Markdown 不在追溯改寫範圍內。
階段載入矩陣
| 階段 | PDCA 階段 | Claude Code 載入 | Gemini 載入 | Codex 載入 |
| Intake | P (pre-Plan) | docs/orchestration.md | -- | -- |
| Research | P (Plan 前準備) | docs/subagent_roles.md §4, docs/subagent_task_templates.md, docs/templates/ | (GEMINI.md has all needed rules) | -- |
| Planning | P (Plan,含 premortem) | docs/schemas/artifact-spec-plan.md §5.3, docs/workflow_state_machine.md, docs/premortem_rules.md | -- | -- |
| Coding | D (Do;微觀內含 TAO) | docs/subagent_roles.md §5, docs/subagent_task_templates.md, docs/templates/ | -- | (CODEX.md has all needed rules) |
| Verification | C (Check) | docs/schemas/artifact-spec-test.md §5.5 + docs/schemas/artifact-spec-verify.md §5.6, docs/workflow_state_machine.md | -- | -- |
| Closure | A (Act;Gate E + 回灌) | docs/workflow_state_machine.md, .github/prompts/remember-capture.prompt.md | Memory Bank Curator 模式時讀 docs/templates/memory-curator/TEMPLATE.md | -- |
| Red Team Exercise | C (Check 衍生) | docs/red_team_runbook.md, docs/red_team_scorecard.md, docs/red_team_backlog.md | -- | -- |
| Sync Contract | meta | docs/orchestration-workflow.md §9 | -- | -- |
兩層架構詳見 docs/orchestration.md §2.8:管理層 PDCA 對應上表「PDCA 階段」欄;執行層 TAO/ReAct 內含於 Coding 階段內 subagent dispatch(詳見 docs/agentic_execution_layer.md)。
交叉引用慣例
- 使用
see docs/X.md §N引用特定章節,避免重複貼內容。 - 範例:"Research artifact format: see
docs/artifact_schema.md§5.2" - Agent 入口檔(CLAUDE/GEMINI/CODEX.md)內嵌了 agent 無法自行額外載入時必須遵守的關鍵規則。
docs/內的參考文件由協調者(Claude Code)依階段按需載入。
章節速查
docs/artifact_schema.md
- §1-§4 通用規則 / §5 索引
docs/schemas/*
- artifact-spec-task.md(§5.1) / artifact-spec-research.md(§5.2) / artifact-spec-plan.md(§5.3)
- artifact-spec-code.md(§5.4) / artifact-spec-test.md(§5.5) / artifact-spec-verify.md(§5.6)
- artifact-spec-decision.md(§5.7) / artifact-spec-status.md(§5.8) / artifact-gallery.md
docs/subagent_roles.md
- §3 Claude Code / §4 Gemini CLI(含 Memory Bank Curator 模式) / §5 Codex CLI / §6 Implementer / §7 Tester / §8 Verifier / §9 Reviewer
docs/premortem_rules.md
- §1-2 When & where / §3 Required fields / §4 Quality rules (P1-P8) / §5 Violation levels / §6 Minimum counts
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

