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 instructions/paceaitian/paceflow/claude-mdgit clone --depth 1 https://github.com/paceaitian/paceflowWhat 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 | $0.01101 | $0.01101 |
| Opus 5 | $0.00550 | $0.00550 |
| Sonnet 5 | $0.00220 | $0.00220 |
| Haiku 4.5 | $0.00110 | $0.00110 |
Grade A, and why
paceflow CLAUDE.md 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 2d 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 — 66 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md
本文件只提供本仓库维护入口。不要把它当作 PACEflow 用户工作流规范，也不要在这里放个人回复风格。
PACEflow 的运行时行为必须由以下位置定义：
- Hook 代码：
plugin/hooks/** - Agent 契约：
plugin/agents/artifact-writer.md与plugin/agent-references/** - 用户 skill：
plugin/skills/**/SKILL.md与其references/** - 发布/安装说明：
README.md、REFERENCE.md
CLAUDE.md 不承担 artifact 创建、批准、验证、归档、Stop gate、worktree owner 或 helper 使用规则的权威职责；这些规则应写回 hooks、skills、agent references 或 README/REFERENCE。
维护规则
- 修改运行时行为时，优先改
plugin/下的实际发布面，并补测试。 docs/、internal/、tests/是仓库维护材料，不随 marketplace runtime 发布。- 历史设计文档保留背景时必须标明 historical / pre-implementation，避免旧 v5 或早期 v6 语义被误当当前规则。
- 不要把用户个人格式要求（时间戳、固定结尾、口癖）写进仓库级
CLAUDE.md。
常用验证
# 一条命令聚合全部核心套件 + claude plugin validate + git diff --check
# 任一子套件非零退出即整体非零退出，杜绝手敲漏跑某套件致回归静默漏网
node tests/run-all.js
# 迭代时按名分片只跑某套件（子串匹配套件名：pace-utils / hooks-e2e /
# session-layers / migrate-v7 / agent-helpers / codex-adapter / mcp-server / run-all-self / plugin-validate / git-diff-check）
PACE_TEST_FILTER=hooks-e2e node tests/run-all.js
# 发版后核整段 release 区间 whitespace：push 后 @{upstream}..HEAD=0，upstream 区间查不到已推送的
# release diff；显式给上一个 release 的 commit 覆盖整段 release 区间 git diff --check（避免 post-push 盲区）
PACE_RELEASE_BASE=<上一个 release 的 commit> node tests/run-all.js
可选本地手动安装检查：如果当前工作区有未跟踪的 install.js 与 tests/test-install.js 本地副本，可以额外运行 node tests/test-install.js。install.js / verify.js 只作为本地 smoke 或手动安装健康检查工具；marketplace 安装以 plugin/.claude-plugin/plugin.json、plugin/hooks/hooks.json 和 plugin/** 发布面为准。
发版
版本号 bump 用 bump-version.js，一条命令同步 5 处，不要手动逐个 Edit（手动易漏某处或改错 changelog 行）：
# 从 paceflow/ 运行（或从仓库根 node paceflow/bump-version.js）
node bump-version.js --dry-run v7.2.16 # 先预览将改哪些文件
node bump-version.js v7.2.16 # 同步 6 处：constants.js PACE_VERSION / .claude-plugin/plugin.json / .codex-plugin/plugin.json / marketplace.json / REFERENCE.md 标题 / README.zh-CN.md 版本行（footer）
脚本不覆盖、仍需手动的两处：
README.md版本历史表格新增一行 changelog（人读描述，脚本不生成）CLAUDE.md等散文里残留的版本号（如有）
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.
- 2d ago First seen · 66 lines · 1,101 tokens per session scan A 2d0df7c2898e
paceflow CLAUDE.md is an instructions file published in the GitHub repository paceaitian/paceflow (38 stars, last pushed 16d ago), licensed MIT. It adds 1,101 tokens to every session, about $0.0055 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 instructions, from other repositories
concord-mcp CLAUDE.md
Instructions for Get-Concord-AI/concord-mcp, covering concord mcp — engineering guide, what this project is, coding rules (non-negotiable), the "never typecast" pattern for db rows and external json and architecture.
eddoapp CLAUDE.md
Instructions for walterra/eddoapp, covering claude.md, build & test commands, commands to run servers, root level and package-specific.
AgentRules AGENTS.md
Instructions for LearnAIHubC/AgentRules, covering agents.md, 1. think before coding, 2. simplicity first, 3. preserve existing architecture and framework boundaries and 4. surgical changes.
pitway AGENTS.md
Instructions for thixpin/pitway, covering agent instructions, project status, what pitway is, architecture constraints (binding on all code) and key operational rules.
agent-context-system AGENTS.md
Instructions for AndreaGriffiths11/agent-context-system, covering agents.md, project, commands, architecture and project knowledge (compressed).
hardproof AGENTS.md
Instructions for asimons81/hardproof, covering hardproof agent instructions, project identity, read order, architecture map and release infrastructure.