MCP HubMCP Hub
← Back to Skills

conscientiousness

pjt222
Updated Yesterday
3 views
17
2
17
View on GitHub
Otherai

About

This skill enables Claude to systematically verify its work, ensuring completeness and accuracy before finalizing a task. It is designed for use when a response needs refinement, after complex multi-step processes, or to correct a pattern of rushing. The core function is to enforce thoroughness by checking that all commitments are met and no corners are cut.

Quick Install

Claude Code

Recommended
Primary
npx skills add pjt222/agent-almanac -a claude-code
Plugin CommandAlternative
/plugin add https://github.com/pjt222/agent-almanac
Git CloneAlternative
git clone https://github.com/pjt222/agent-almanac.git ~/.claude/skills/conscientiousness

Copy and paste this command in Claude Code to install this skill

Documentation

盡責

系統之周詳與勤——確完整、驗結果、行其諾、完任務至其當得之標。

適用時機

  • 標任務為完前——為最終驗證之過
  • 回應覺「足好」然任務當得更佳
  • 複雜多步作業後，各步或有偏
  • 用戶請求含多部，每部需驗
  • 呈代碼、文件、任何交付物予用戶審前
  • 自監偵省走捷徑或匆忙之模式

輸入

  • 必要：待驗之任務或交付物（於對話脈絡可達）
  • 選擇性：用戶原始請求（以比所交付）
  • 選擇性：用戶供之檢核表或受理判準
  • 選擇性：會話中所作之承諾（已諾而未查者）

步驟

步驟一：重構全承諾

驗工作前，重立所諾為何。

  1. 仔讀用戶原始請求——非詮之版本，乃實之語
  2. 列其中每一明示要求
  3. 列會話中每一隱含承諾：
    • 「我亦將更新測試」——已為之否？
    • 「讓我亦修此」——已完否？
    • 「我將查邊界情形」——已查否？
  4. 記用戶供之受理判準
  5. 比諾單與實交付之

預期： 完整之承諾清單——明示要求加隱含之諾——附與交付物之初步比。

失敗時： 若原始請求已不在脈絡（已壓縮），自所餘重構並向用戶承認任何缺。

步驟二：驗完整

查每一已諾項已處。

Completeness Matrix:
+---------------------+------------------+------------------+
| Commitment          | Status           | Evidence         |
+---------------------+------------------+------------------+
| [Requirement 1]     | Done / Partial / | [How verified]   |
|                     | Missing          |                  |
+---------------------+------------------+------------------+
| [Requirement 2]     | Done / Partial / | [How verified]   |
|                     | Missing          |                  |
+---------------------+------------------+------------------+
| [Promise 1]         | Done / Partial / | [How verified]   |
|                     | Missing          |                  |
+---------------------+------------------+------------------+
  1. 每項以證驗——非記憶，乃實之驗：
    • 代碼改：重讀檔以確改存
    • 測試結果：重行或引實輸出
    • 文件：重讀以確準
  2. 標每項：Done（全完）、Partial（始而未完）、Missing（未處）
  3. Partial 與 Missing 項記所餘

預期： 每承諾有已驗之態。無項未查。

失敗時： 若驗揭漏項，即處勿記後。盡責乃即完，非擬完。

步驟三：驗正確

完整為必須而不足——所為須亦正確。

  1. 每已完項查：
    • 準：行其當行？值正確？
    • 一致：合其餘之工作？無矛盾？
    • 邊界情形：邊界條件已慮？
    • 整合：於周圍脈絡中可運？
  2. 代碼：能存於代碼審否？有明改進否？
  3. 文件：準、清、無誤否？
  4. 多步流程：每步之出正確餵次步否？

預期： 每交付物既完且正。錯於用戶見前已捕。

失敗時： 若發錯，即修。勿以已知錯之工呈，即錯似微。

步驟四：驗呈現

最終檢查：交付物之呈現服用戶否？

  1. 清：用戶無需重讀多遍即解所為否？
  2. 組：回應結構合邏輯？相關項已群？
  3. 簡：有無謂填充或重複否？
  4. 可行：用戶知下步何為否？
  5. 誠：限制或警語已明述否？

預期： 交付物完、正、良呈。

失敗時： 若內容正而呈現拙，重組。良工拙呈乃盡責之敗。

驗證

  • 原始請求已重讀（非自記憶召回）
  • 每明示要求以證驗
  • 每隱含之諾已追且驗
  • 正確已查，超於僅完整
  • 相關時邊界情形已慮
  • 交付物清晰呈現且可行

常見陷阱

  • 驗證戲：行檢查之動作而未實重讀或重驗。檢查須用證，非記憶
  • 局部盡責：查主交付物而忽附承諾（「我亦將⋯」）。每諾皆計
  • 偽裝勤之完美主義：無盡磨延交付。盡責乃達所諾之標，非無限超之
  • 盡責疲勞：會話進行中轉不周。末任務當得之勤與初同
  • 簡單任務略之：假簡單任務不需驗。簡單任務之錯較複雜任務之錯更窘

相關技能

  • honesty-humility — 盡責驗完整；honesty-humility 確透明報何者成何者未成
  • heal — 子系統評估與自驗重疊；盡責專於交付物品質
  • vishnu-bhaga — 保可行態補盡責於維品質
  • observe — 持續中立觀察支驗證過程
  • intrinsic — 真誠投入（非從順）自然驅周詳執行

GitHub Repository

pjt222/agent-almanac
Path: i18n/wenyan-lite/skills/conscientiousness
0
agentsagentskillsai-assisted-developmentclaude-codeskillsteams

Related Skills

llamaguard

Other

LlamaGuard is Meta's 7-8B parameter model for moderating LLM inputs and outputs across six safety categories like violence and hate speech. It offers 94-95% accuracy and can be deployed using vLLM, Hugging Face, or Amazon SageMaker. Use this skill to easily integrate content filtering and safety guardrails into your AI applications.

View skill

cost-optimization

Other

This Claude Skill helps developers optimize cloud costs through resource rightsizing, tagging strategies, and spending analysis. It provides a framework for reducing cloud expenses and implementing cost governance across AWS, Azure, and GCP. Use it when you need to analyze infrastructure costs, right-size resources, or meet budget constraints.

View skill

quantizing-models-bitsandbytes

Other

This skill quantizes LLMs to 8-bit or 4-bit precision using bitsandbytes, achieving 50-75% memory reduction with minimal accuracy loss. It's ideal for running larger models on limited GPU memory or accelerating inference, supporting formats like INT8, NF4, and FP4. The skill integrates with HuggingFace Transformers and enables QLoRA training and 8-bit optimizers.

View skill

dispatching-parallel-agents

Other

This Claude Skill dispatches multiple agents to investigate and fix 3+ independent problems concurrently. It is designed for scenarios involving unrelated failures that can be resolved without shared state or dependencies. The core capability is parallel problem-solving, assigning one agent per independent problem domain to maximize efficiency.

View skill