gtd-clarify

gtd-clarify is a skill for Claude Code from mikonos/LLM-GTD. It costs 69 tokens per session (2,317 once invoked), scanned A, original, MIT.

A GTD tool that processes items in an inbox by deciding whether they require action and where they belong. It turns vague entries into specific next actions, delegated tasks, calendar items, projects, reference material, or future possibilities.

In plain words
What is it for?
Use it to clear an inbox, classify imported tasks, define the next physical step, delegate work, create projects, and place time-specific commitments on a calendar when possible.
Why use it?
It removes the uncertainty of having an inbox full of unprocessed notes and reminders. The decision process helps separate actionable work from information, discarded items, and things to consider later.

Skill for Claude Code

Written for Claude Code: shipped in a Claude Code plugin.

Part of the llm-gtd plugin — 8 skills, 8 commands shipped together

Good fit Use it to clear an inbox, classify imported tasks, define the next physical step, delegate work, create projects, and place time-specific commitments on a calendar when possible.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/mikonos/llm-gtd/clarify
Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

Any agent
npx skills add mikonos/LLM-GTD --skill clarify
Clone the repo
git clone --depth 1 https://github.com/mikonos/LLM-GTD

Made for: Claude Code.

Or install llm-gtd, the plugin that ships this one along with the rest of its 8 skills, 8 commands.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for gtd-clarify

README.md
[![agentmods](https://agentmods.dev/badge/skills/mikonos/llm-gtd/clarify/github.svg)](https://agentmods.dev/skills/mikonos/llm-gtd/clarify)
Your own site
<a href="https://agentmods.dev/skills/mikonos/llm-gtd/clarify"><img src="https://agentmods.dev/badge/skills/mikonos/llm-gtd/clarify/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for gtd-clarify

Your own site · 80×15
<a href="https://agentmods.dev/skills/mikonos/llm-gtd/clarify"><img src="https://agentmods.dev/badge/skills/mikonos/llm-gtd/clarify.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 69 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 2,317 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00069 $0.02317
Opus 5 $0.00034 $0.01158
Sonnet 5 $0.00014 $0.00463
Haiku 4.5 $0.00007 $0.00232

Measured 9d ago against content hash 29c01d4aa6c1, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-09, from the pricing page.

Security

Grade A, and why

gtd-clarify scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 9d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

plugins/llm-gtd/skills/gtd-harness/clarify/SKILL.md · 81 lines

How it starts

The opening of the file, as written. The whole thing — 81 lines — stays where its author put it; the contents beside it link to each section on GitHub.

GTD · clarify(理清 + 归位)

视角:David Allen。Clarify 是 GTD 最锋利的工序——把模糊的「东西(stuff)」逐项问成可执行的结论。多数人的待办系统瘫痪,就因为收件箱里堆的是没被理清的「东西」,而不是「下一步行动」。理清和归位(organize)通常一气呵成,故本命令兼做落位。

加载与边界

  • 清单去向和动作权限:先读 references/list-definitions.md
  • 下一步含糊或需要解释标准:读 references/clarify-decision-tree.md
  • 涉及 hard landscape / 写日历:读 references/capability-map.md
  • 自动化边界:清楚的条目可自动归位;承诺本身、最终完成口径或外部写入冲突不清时问一句。

何时跑

  • inbox.md 有待理清项(dashboard 提示)。
  • 用户说「帮我理一理收件箱」「这些待办怎么归」「逐条过一遍」。
  • 导入旧数据后「待补轻字段」段需要重判。

决策树(对 inbox 每一项逐条问)

这是什么?它需要行动吗?
├── 否 → 三个去向:
│   ├── 没用/过期        → 删除(垃圾桶)
│   ├── 暂不做但不愿忘   → someday-maybe.md
│   ├── 产品/功能/场景机会 → product-ideas.md(保留机会原文)+ projects/next-actions(日常可见)
│   └── 备查/支持材料    → reference.md(知识/想法类 → 移交 ZK 管线 fleeting-note)
│
└── 是 → 下一步具体的物理动作是什么?
    ├── < 2 分钟         → 立刻做(two-minute rule),做完即销项
    ├── 该别人做         → 委派 → waiting-for.md(记人名+约定+日期)
    └── 自己做、>2 分钟  →
         ├── 特定时间/日才做  → hard landscape(日历):先查目标时间窗;无冲突且信息完整时写入可达的 calendar provider;不可达或信息缺关键字段则降级
         └── 尽快做          → next-actions.md(行动池;写清预计时长 / 精力档 / 真实约束)
    ※ 若完成需要 >1 步 → 同时在 projects.md 立项(成果+下一步),下一步进 next-actions

工作流

  1. inbox.md,逐项跑决策树。一次理清一项,不跳。
  2. 关键问题
    • 「需要行动吗?」——这道闸把行动与知识分开(知识 → ZK 管线,不进 GTD 清单)。
    • 「下一步具体的物理动作是什么?」——必须是可见的物理动作("打电话给同事 A 确认资料口径"),不是"处理资料"。动词具体,拒绝空动词。
    • 「这条的完成到底算什么?」——先定义 outcome / 完成口径,再决定它该进 waiting-fornext-actionsprojects 还是可以闭环。不要把“approval passed”“约上时间”“拿到回复”这类里程碑,自动当成最终完成,除非用户明确这样定义。
    • 「这条行动的预计时长 / 精力档 / 真实约束是什么?」——只写轻字段,不把镜头变成复杂标签系统。默认由 AI 估计,明显拿不准才问一句。
  3. 落位:按决策树把该项追加到目标清单对应分组末尾(格式见各清单模板)。
    • 进 next-actions 必带三类轻字段:预计时长(2分钟 / 10分钟 / 30分钟 / 60-90分钟)、精力档(低精力 / 中精力 / 深工作 / 情绪耗能低)、真实约束(硬地点/场景、工具/渠道、人在场、准备链、采购、会议前、证件/付款/文件/设备等)。旧 @电脑/@电话/@外出/@家/@议程-人 分组仅作兼容,不再默认当主分类。
    • 明确是产品想法/功能机会/场景机会时,进 product-ideas.md 保留原始机会,并默认同步创建 projects.md + next-actions.md 的可见工作项;只有用户明确说「先存不处理 / 只捕捉」才暂不升级。
    • 立项时在 projects.md 写成果 + 下一步,并把下一步同步进 next-actions。项目里的「下一步行动」不要只写「见 next-actions」或泛泛情境;应链接到 next-actions 的具体条目 block,例如先在 action 行末加 ^na-short-id-YYYYMMDD,再在 project 写 [[next-actions#^na-short-id-YYYYMMDD|具体下一步行动]](约束:需要电脑)
    • GTD 内部标题引用用 Obsidian heading link:项目引用写 [[projects#项目名|项目名]];支持材料引用写 [[reference#条目名|条目名]]。裸 [[标题]] 只用于真实独立文件。
    • 更新既有项目时,若期望成果已经达成且没有仍需推动的下一步 → 从 projects.md 删除整个项目块;不要写「下一步行动:无」或把已完成项目留作记录。
    • 等待项(waiting-for)要同时写清两层当前在等什么最终什么才算完成。尤其是 payment flow、approval flow、法务流转、退款、返款这类事项,approval passed / 已提交 / 对方回复 / 技术评估开始 往往只是里程碑,不等于闭环。若最终完成口径是“到账 / 签字完成 / 实物收到 / 真正约成并发生”,就必须在 waiting-for 文案或约定里写明,避免提前删项。
    • 2 分钟能做完的,提示用户「这条 <2 分钟,建议现在就做」。
    • 特定时间事 → 日历(hard landscape):单一日历铁律——外部 calendar provider 可达时优先写外部 provider,全部不可达才写 calendar.md 兜底;不抄副本。写外部日历是高后果操作,但日程信息完整时可自动写入;写入前先读目标时间段的 hard landscape,若同时间段已有事件或前后缓冲不足以履约,列出冲突并停止写入,提示改期、取消、委派或降级为 next-action/someday;外部 tool 返回成功才报「已写入」,失败/不可达则按 references/capability-map.md 逐级降级并如实说明,绝不谎报。缺日期、时间、标题/对象等关键字段时,只问缺失字段;会议缺时长时默认 60 分钟。
  4. 落位后从 inbox 删除该项(理清完就不该再留收件箱)。
  5. 信息不足以判断下一步 → 落 next-actions 写「TBD + 缺什么信息」,或反问用户一句。

Read the full file on GitHub · 81 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 9d ago First seen · 81 lines · 69 tokens per session scan A 29c01d4aa6c1

Subscribe to this mod's changes

gtd-clarify is a skill published in the GitHub repository mikonos/LLM-GTD (10 stars, last pushed 3mo ago), licensed MIT. It adds 69 tokens to every session and 2,317 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.