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.
npx agentmods add skills/backtocimacoppi/praxis/task-control-docnpx skills add BackToCimaCoppi/Praxis --skill task-control-docgit clone --depth 1 https://github.com/BackToCimaCoppi/PraxisWrote 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/backtocimacoppi/praxis/task-control-doc)<a href="https://agentmods.dev/skills/backtocimacoppi/praxis/task-control-doc"><img src="https://agentmods.dev/badge/skills/backtocimacoppi/praxis/task-control-doc.svg" alt="Measured on agentmods" height="20"></a>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 | $0.00066 | $0.08315 |
| Opus 5 | $0.00033 | $0.04158 |
| Sonnet 5 | $0.00013 | $0.01663 |
| Haiku 4.5 | $0.00007 | $0.00831 |
Grade A, and why
task-control-doc 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 3d 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 — 456 lines — stays where its author put it; the contents beside it link to each section on GitHub.
任务总控文档
当用户要求"为某件事做总控文档"时,使用本 skill。
真值源:~/.claude/skills/control/references/总控规范.md。本 skill 只描述如何创建总控;生命周期、归档定义、目录结构都在那里。
方法论补充:references/方法论.md(为什么要做、何时做、常见风险)。
1. 适用场景
- 任务很大、很复杂
- 任务可能跨多个会话完成
- 用户希望每个子任务都开新会话执行(这是默认假设)
- 上下文可能过长、容易污染或遗忘
中小任务用计划模式或 lightweight-design 即可,不需要总控。
2. 核心原则:子任务即工作包
这是本 skill 最重要的一条设计原则。
每个子任务都应该是一个自包含的工作包:
- 新会话读「子任务详情 + 强制阅读文件」即可开工——这是准入下限,不是视野上限
- 鼓励执行会话开工前主动补读:「背景导航」列出的文件、父级总控、其他子任务详情、相关正式文档、代码现状——把背景挖够再动手。强模型(Fable 5 / GPT-5.6 级)能自主取舍读什么;背景不足导致误判的代价,远大于多读几个文件
- 视野放开、扇出焊死:自主补读 = 亲自读(直接读文件 / 检索),禁止为"补背景"派子 agent / 起深度调查(读是加法、派 agent 是乘法;执行会话对背景问题是叶子)。觉得背景缺口大到需要专门调查 → 说明工作包本身没写清,停下向用户报告
- 执行与写入范围仍严格限于本子任务——读什么放开 ≠ 做什么放开
- 每个工作包必须记录
start_commit、allowed_write_paths、allowed_cross_task_writes;交付前对本子任务明确列出的候选提交运行control/scripts/check_write_scope.py。无关提交不参与本子任务检查,候选提交夹带越界文件则直接失败 - 用户在新会话开头自己选模型,不预定义执行模式
- 子任务详情末尾有「会话启动提示词」可直接复制
写总控时按这个原则切分子任务:强制阅读(核心必读)精准 1-3 个(超了说明背景没消化成任务),背景导航不设上限(一行一条「路径 + 读它获得什么」,宁多勿缺)。
标准研发总控的强制阅读按职责放置,禁止只在背景导航里“提到但不读”:测试用例阶段绑定 test-standards + test-case-design;goal 章程阶段绑定 goal-charter;goal 执行阶段绑定章程、test-execution-router 与项目执行/部署/架构入口;候选终审阶段绑定证据协议、项目证据规则与冻结 L7。具体项目名由研发流程补丁填充。
3. 文件位置与结构
默认路径:<PROJECT_ROOT>/docs/00-任务总控/{YYYY-MM-DD}-{任务名}/
任务目录用日期前缀(创建日,YYYY-MM-DD),防多 worktree 编号撞车。同日创建多任务可加字母后缀
2026-05-10b-...或时分2026-05-10-1430-...。
支持两种模式(创建时选择):
3.1 单文件模式(默认,适合中小总控)
{YYYY-MM-DD}-{任务名}/
├── README.md # 主总控(含全部子任务详情)
└── _shared/ # 可选:唯一过程资产目录(归属由文件名前缀区分,见 §3.3)
适用:3-7 个子任务、各子任务详情不超过 50 行。
3.2 拆分模式(适合大总控)
{YYYY-MM-DD}-{任务名}/
├── README.md # 主总控(任务背景 + 子任务总表 + 进展记录,不含子任务详情)
├── T1-{子任务名}.md # T1 自包含工作包
├── T2-{子任务名}.md # T2 自包含工作包
├── T6-{子任务名}.md
└── _shared/ # 可选:唯一过程资产目录(归属由文件名前缀区分,见 §3.3)
├── T3-{资产文件}.csv # 子任务专属资产
└── {资产文件}.md # 任务级共享资产(无前缀)
适用:8+ 个子任务、每个子任务详情很长(带大量强制阅读、设计明细)。
What ships with it
18 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.
- assets/任务总控模板.md 6.3 KB
- assets/拆分模板/README.md 2.9 KB
- assets/拆分模板/子任务包.md 4.3 KB
- assets/研发流程模板/README.md 4.1 KB
- assets/研发流程模板/规格物化覆盖报告模板.json 3.7 KB
- assets/研发流程模板/规格物化覆盖报告模板.md 1.9 KB
- assets/研发流程模板/阶段片段库/goal执行.md 11 KB
- assets/研发流程模板/阶段片段库/goal章程.md 10 KB
- assets/研发流程模板/阶段片段库/候选终审.md 5.9 KB
- assets/研发流程模板/阶段片段库/开放探讨.md 3.0 KB
- assets/研发流程模板/阶段片段库/测试用例设计.md 5.3 KB
- assets/研发流程模板/阶段片段库/规格冻结.md 7.1 KB
- assets/研发流程模板/阶段片段库/评审.md 4.7 KB
- assets/研发流程模板/阶段片段库/调查.md 3.5 KB
- assets/研发流程模板/阶段片段库/轻量设计.md 4.5 KB
- references/方法论.md 8.2 KB
- references/标准研发流程.md 32 KB
- tests/test_standard_flow_templates.py 4.0 KB runs code
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.
- 3d ago First seen · 456 lines · 66 tokens per session scan A be5cfaa9e4bb
task-control-doc is a skill published in the GitHub repository BackToCimaCoppi/Praxis (7 stars, last pushed 9d ago), licensed Apache-2.0. It adds 66 tokens to every session and 8,315 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.
Other skills, from other repositories
neurolink-guide
Guide for using the NeuroLink SDK and CLI. Invoke when users ask how to use neurolink, integrate AI providers, add MCP tools, configure RAG, set up memory, deploy servers, or work with multimodal content. Covers SDK, CLI, providers, tools, and enterprise features.
data-validator
独立于主流程的数据校验 agent。物理只读、无写权限、独立加载规则——避免「自己写自己判」的自评偏差。 当以下情况时使用: (1) 写完 SQL,跑之前需要校验 (2) 数据拉完,写入报表/文档前需要校验 (3) 分析报告写完,交付前需要校验 (4) 用户说「校验一下」「检查一下」「再 review 一遍」.
critical-review
Radical-honesty architectural review — questions code, methodology, design, and operational fitness.
systematic-debugging
Root-cause a bug with a hypothesis-driven loop instead of shotgun edits. Use when a bug isn't obvious after the first look, or when asked to "find out why" something fails.
triz-synergy
Разрешает противоречия в разработке: формулирует пару взаимоисключающих требований к одному объекту, ищет уже существующий ресурс в коде и в структуре данных, разделяет по структуре/времени/условию/отношению и проверяет решение различающим опытом. On-demand only — вызывать явно, когда виден признак противоречия. Use…
code-writing
Universal quality coding process: plan, TDD, reviews. Use whenever code needs to be written — ad-hoc or as part of a task. Use when: "напиши код", "закодь", "реализуй", "write code", "implement" For planning tasks → tech-spec-planning skill. For specs → user-spec-planning skill.