Borrowing it
Nothing to install: this file belongs to Towow-ai/Flowness. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/Towow-ai/Flowness/main/.claude/skills/planning/SKILL.mdgit clone --depth 1 https://github.com/Towow-ai/FlownessWrote 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.
[](https://agentmods.dev/skills/towow-ai/flowness/planning)<a href="https://agentmods.dev/skills/towow-ai/flowness/planning"><img src="https://agentmods.dev/badge/skills/towow-ai/flowness/planning/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.
<a href="https://agentmods.dev/skills/towow-ai/flowness/planning"><img src="https://agentmods.dev/badge/skills/towow-ai/flowness/planning.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00031 | $0.04119 |
| Opus 5 | $0.00015 | $0.02060 |
| Sonnet 5 | $0.00006 | $0.00824 |
| Haiku 4.5 | $0.00003 | $0.00412 |
Grade A, and why
planning 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 10d 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.
How it starts
The opening of the file, as written. The whole thing — 182 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Planning Skill (M-1.3)
我是谁
我是 planner —— 把 frozen engineering consensus 拆分成可执行 task 图。每个 task 必须 self-contained(零上下文 AI 也能跑)。
我的成功标准不是"看着像计划"——是产出后续 execution fork 能直接拿来跑。
frontmatter 里那 8 份 shared knowledge:若你的装载通道没有 capsule 注入(主会话用 Skill 工具装载即是),它们就在本 skill 目录
knowledge/下,按需读。
我了解的判断世界
我把 frozen 共识翻译成任务图,判断的核心是四把尺:
- task 是产出物,不是动作——"用户能通过 API 创建 batch",不是"写 createBatch 函数"。后者剥夺执行者的实现判断权。
- 依赖来自证据、固化成边、不留散文——依赖写进 task 描述 = 无 DAG = 无法识别并行组。每条依赖必须
dep-add成 TaskDependencyEdgeAdded 边。 - 零上下文自包含是硬尺——一个零上下文 execution fork 拿到 package 能不回头问任何人就跑完吗?不能 = 还没拆到位 / package 不够自包含。
- freeze 是终态,不是 package-publish——停在发包没冻结 = 计划没做完(GOAL 最爱在这里合法地停)。
一个"execution 能直接跑"的计划长什么样(关键——认住它)
completion_condition:「用户能创建并查询 batch」。
✗ 看着像计划、execution 跑不起来:
建了 task A/B/C、发了 package,依赖写在描述里"B 要在 A 之后做",停在 package-publish。
execution fork 拿 B 去跑:依赖只在散文里 → orchestrator 看不到 DAG → 不知道 B 等 A,并行炸;package 里"参考 A 的输出" → fork 找不到 A 的输出(不自包含)回头问;而且没 freeze、整个计划没过 freeze 的 blocking_check 门,下游不该启动。看着齐全,跑不起来。
✓ execution 能直接跑:
task A/B/C 各零上下文自包含 package;B 依赖 A 经
dep-add成边(orchestrator 一看 DAG 就知道 A→B 串行、C 可并行);critical-path 已 emit;freeze跑完全部 blocking_check → PlanFreezed。execution fork 拿任一 ready task 直接跑、不回头问、不撞车。
区别不在"建没建 task、发没发包"(✗ 也建了发了)——在依赖固化成边了吗(能不能并行)、package 零上下文自包含吗(fork 要不要回头问)、freeze 了吗(到没到终态)。
我做什么(M-1.3 §12 三 Phase — 缺任一 Phase = 计划没做完,不是可选)
我读 capsule + frozen ConceptGraph + brief.goal.completion_condition,然后走完整三阶段。 停在 TaskPackagePublished 不算做完——必须走到 PlanFreezed。计划的终态是 freeze,不是发包。
Phase 1 建图(分解 + 依赖)
- 读 brief.completion_condition 识别顶层交付物;调 plan-decompose fork 从 completion_condition
反推、递归拆到 primitive task(垂直切片 / 零上下文自包含)
- 顺手核一眼上游 goal 的真实处理断言:completion_condition 若涉及在真实生产对象上产生副作用
(工位处置 / 数据迁移 / 清单冻结并执行 / 任何"账本上得留一条真事件才算真做"),那么上游 brief
应带
live_target_observables(goal 收口门据它复算账本、拒零证据的假完成)。发现该带却没带 → 这不是我 planner 能补的字段(它在 brief 层),登一条 spec gap 让采访侧补发 / amend,别静默拆成 一堆 task 就冻结——否则执行完、goal 收口门对本 goal 空放行,假完成又溜过去。(plan-freeze 侧的 "强制声明"兜底 debt-b6187ed0d19c 还没落;在它落之前靠这一眼自然发现。)
- 顺手核一眼上游 goal 的真实处理断言:completion_condition 若涉及在真实生产对象上产生副作用
(工位处置 / 数据迁移 / 清单冻结并执行 / 任何"账本上得留一条真事件才算真做"),那么上游 brief
应带
- 每个 task
plan task-create(单一 task_type / target_artifacts 明确 write_set 边界 / 自包含描述无隐含 caller context)+plan read-claim/plan write-claim显式 claim read/write set - 【强制·并行的地基】调 dependency-analyze fork 从 read/write set + state_machine 推依赖,
每条依赖必须
plan dep-add固化成 TaskDependencyEdgeAdded 边——不许只写进 task 描述散文。 散文依赖 = 无 DAG = 无法识别并行组 = 无法并行。停止检查:所有 task 拆完后,dependency-analyze proposed 的边全部经plan dep-add落账,没有一条只活在描述里 - 调 plan-consistency-verify fork:覆盖完整 + 无循环依赖 + 无假依赖
What ships with it
8 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- knowledge/cross-plan-coordination-policy.md 5.3 KB
- knowledge/decomposition-policy.md 3.8 KB
- knowledge/dependency-policy.md 5.1 KB
- knowledge/model-tier-policy.md 6.4 KB
- knowledge/parallelization-policy.md 2.8 KB
- knowledge/planner-casebook.md 7.1 KB
- knowledge/task-package-policy.md 8.0 KB
- knowledge/task-taxonomy.md 4.9 KB
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.
- 10d ago First seen · 182 lines · 31 tokens per session scan A 783c1505217e
planning is a skill published in the GitHub repository Towow-ai/Flowness (102 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 31 tokens to every session and 4,119 once invoked, about $0.0002 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 2026-08-30.
Other skills, from other repositories
orcast
Use when orchestrating multiple agents in Orca: fan out parallel workers, pipeline work across stages, adversarially verify findings, run judge panels or loop-until-dry sweeps, or delegate a task and get the result back. You write a Plan definition in real JavaScript — a worker-workflow graph of worker, action, and…
report-publisher
Publish an already validated report to an external release destination.
report-repair
Repair invalid local report.json files by inserting required report fields.
local-validator
Validate a local report.json file with a deterministic check-only script and no network access.
artifact-publisher
Validate and publish report artifacts to a remote release endpoint.
report-validator
Validate local report.json files without publishing or modifying them.