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 skills add RainFlashPoint/capital-agent-skills --skill cap-flowgit clone --depth 1 https://github.com/RainFlashPoint/capital-agent-skillsWrote 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/rainflashpoint/capital-agent-skills/cap-flow)<a href="https://agentmods.dev/skills/rainflashpoint/capital-agent-skills/cap-flow"><img src="https://agentmods.dev/badge/skills/rainflashpoint/capital-agent-skills/cap-flow.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 5.1 | $0.00444 | $0.09906 |
| Opus 5 | $0.00222 | $0.04953 |
| Sonnet 5 | $0.00089 | $0.01981 |
| Haiku 4.5 | $0.00044 | $0.00991 |
Grade A, and why
cap-flow 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 yesterday.
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 — 531 lines — stays where its author put it; the contents beside it link to each section on GitHub.
cap-flow — 主线编排器（导演 / 路由 / 交接）
两条全程硬契约
每次进入后先加载并执行：
references/progress-protocol.md：阶段开始、执行中和交接都必须向用户明确说明当前动作与下一步。references/task-reconnaissance.md：每个新任务必须基于当前仓库真实代码建立或刷新.cap/task-context.md；PROFILE.md只作索引，不能替代任务级代码调查。references/harness-action-protocol.md：Test/Review/Patch 的可信执行边界；STATE 是游标，不是 Gate 证明。
任一阶段被直接调用时也必须遵守这两条契约。没有新鲜 task-context.md 时，不得直接进入计划或编码。
准备进入 plan / implement / test / review / release 时，必须运行确定性 Context Guard；POSIX 沿用 scripts/cap-context-guard，原生 Windows 通过 package 根 scripts/cap-runtime.mjs context 执行同一契约。失败就留在当前阶段刷新代码调查，不能只靠模型自判“应该没问题”。
第三条硬契约：上下文预算（单会话多阶段强制）
持久真相在 .cap/ 纯文件里，不在会话上下文里——所以每过一个阶段就回收上下文，绝不只增不减：
- 按需读要点：reference / playbook / 角色卡只取当前决策需要的段落，不无脑读全文；确需全文才读全文。
- 角色卡一次一张：进 implement / review 时一次只装载一张活跃角色卡，用完即弃，绝不预载全部角色。
- 阶段间回收：写完
## HANDOFF并落 STATE 后，该阶段的临时输入（角色卡、大 reference、原始 diff / 证据）视为已丢弃；下一阶段只从.cap/STATE.md+ 真正需要的文件重建最小上下文，绝不把上一阶段的上下文整包带过去。 - 读过即在，不重复读盘：本会话已读入的文件（SKILL / reference / 角色卡 /
.cap/*）默认已在上下文。后文（及各阶段）反复出现的“先加载并执行 X”“各阶段沿用”只表示“确保该纪律生效”，不等于重新读盘：若该文件本会话已读且磁盘未变，直接沿用已读内容，不得再次read/sed读第二遍；只有从未读过、或文件确已变更，才真正读盘。
有子 agent 隔离（Claude）就 fan-out 让大材料留在子上下文；无隔离（如 Codex）时上述纪律强制执行（serial-and-evict），否则单会话累积到数百 KB 会让下一轮请求在上游网关 504 超时。展开见文末「上下文预算」与 references/runtime-adapters/codex.md 的 Context budget 段。
公开语言与内部 ID
研发只需要记住 $cap（Claude Code 为 /cap）。所有对话、选项、下一步和错误提示都优先使用下列直白名称，不得要求用户调用内部技能名：
| 对外名称 | 可选快捷表达 | 内部状态 / 技能 |
|---|---|---|
| 项目了解 | /cap 项目 |
understand / cap-understand |
| 需求确认 | /cap 需求 |
define / cap-define |
| 开发计划 | /cap 计划 |
plan / cap-plan |
| 编码实现 | /cap 开发 |
implement / cap-implement |
| 测试验证 | /cap 测试 |
test / cap-test |
| 代码评审 | /cap 评审 |
review / cap-review |
| 发布上线 | /cap 发布 |
release / cap-release |
内部文件、路由表和兼容调用继续使用稳定 ID，避免破坏历史 .cap/STATE.md、平台事件与旧客户端。若必须展示内部 ID，只能作为诊断补充，例如“当前：测试验证（内部阶段 verify）”。
你在这条研发主线里扮演导演,不亲自演任何一场戏。你的全部工作围绕三个动作展开:
Orient(定位) → Route(派发) → Handoff(交接)
读清现在在哪 决定去哪、带什么 把结果记下来、告诉用户下一步
What ships with it
45 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.
- references/build-loop.md 15 KB
- references/collaboration-discipline.md 14 KB
- references/complexity-routing.md 3.4 KB
- references/cross-project-handoff.md 3.4 KB
- references/deployment-patterns.md 18 KB
- references/distillation-loop.md 14 KB
- references/divergence-frames.md 5.6 KB
- references/evolve-loop.md 7.3 KB
- references/harness-action-protocol.md 8.9 KB
- references/headless-policy.md 6.4 KB
- references/independent-review.md 4.8 KB
- references/intake.md 20 KB
- references/languages/go.md 6.5 KB
- references/languages/java-spring.md 6.4 KB
- references/languages/kotlin.md 8.6 KB
- references/languages/python.md 8.2 KB
- references/languages/rust.md 8.1 KB
- references/languages/swift.md 9.1 KB
- references/languages/typescript.md 6.4 KB
- references/progress-protocol.md 8.5 KB
- references/receiving-feedback.md 2.6 KB
- references/role-routing.md 13 KB
- references/roles/ai-readiness.md 5.3 KB
- references/roles/architect.md 4.9 KB
- references/roles/big-data.md 12 KB
- references/roles/client-dev.md 10 KB
- references/roles/design.md 13 KB
- references/roles/qa.md 12 KB
- references/roles/server-dev.md 11 KB
- references/roles/skill-maintainer.md 7.1 KB
- references/runtime-adapters/codex.md 11 KB
- references/task-reconnaissance.md 6.8 KB
- references/templates/hooks/post-checkout 2.5 KB
- references/templates/hooks/pre-commit 1.6 KB
- references/templates/hooks/pre-push 4.9 KB
- references/templates/PROFILE.md 7.4 KB
- references/templates/STATE.md 11 KB
- references/web-review/annotate.css 4.9 KB
- references/web-review/annotate.js 11 KB runs code
- references/web-review/build.py 8.3 KB runs code
- references/web-review/playbook.md 7.4 KB
- references/web-review/server.py 3.8 KB runs code
- references/web-review/test_live.py 4.2 KB runs code
- scripts/cap-context-guard 5.0 KB
- scripts/cap-guard 2.9 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.
- yesterday Changed · +10 lines 17b28b3eb9ea
- 3d ago Changed 2864765d5837
- 6d ago First seen · 521 lines · 444 tokens per session scan A 2ed90333a120
cap-flow is a skill published in the GitHub repository RainFlashPoint/capital-agent-skills (2 stars, last pushed 2d ago), licensed MIT. It adds 444 tokens to every session and 9,906 once invoked, about $0.0022 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
systematic-debugging
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.
local-ai-agents
Build local-first AI agents that run entirely on a developer workstation with Microsoft Foundry Local and Qwen function-calling models. Covers Small Language Models (SLMs), the OpenAI-compatible local endpoint, sandboxed local tools, local RAG with Chroma, local MCP servers, hybrid cloud/local routing, and the…
next-cache-components-adoption
Turn on Cache Components in a Next.js app and resolve the blocking routes it surfaces. Use when the user wants to enable, adopt, or migrate to Cache Components, flip the cacheComponents flag, work through a flood of blocking-prerender / instant validation errors, run the cache-components-instant-false codemod, or…
next-partial-prefetching-adoption
Turn on Partial Prefetching in a Next.js app and work through the insights it surfaces. Use when the user wants to enable or adopt Partial Prefetching, flip the partialPrefetching flag, opt routes in with export const prefetch = 'partial', audit Link prefetch={true} behavior, preserve existing prefetched UI with…
chronicle
Analyze Copilot session history for standup reports, usage tips, session search, and session reindexing. Use when the user asks for a standup, daily summary, usage tips, workflow recommendations, wants to search or find past sessions by keyword/file/PR, wants to reindex their session store, or asks about deleting…
babysit-pr
Babysit a GitHub pull request after creation by continuously polling review comments, CI checks/workflow runs, and mergeability state until the PR is merged/closed or user help is required. Diagnose failures, retry likely flaky failures up to 3 times, auto-fix/push branch-related issues when appropriate, and keep…