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/gzchenhao/openhire/claude-mdgit clone --depth 1 https://github.com/gzchenhao/openhireWrote 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/instructions/gzchenhao/openhire/claude-md)<a href="https://agentmods.dev/instructions/gzchenhao/openhire/claude-md"><img src="https://agentmods.dev/badge/instructions/gzchenhao/openhire/claude-md.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 | $0.01944 | $0.01944 |
| Opus 5 | $0.00972 | $0.00972 |
| Sonnet 5 | $0.00389 | $0.00389 |
| Haiku 4.5 | $0.00194 | $0.00194 |
Grade A, and why
openhire 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 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 — 66 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md — OpenHire 工作约定(每个会话必读)
新会话恢复指引(开工前先做这三步)
- 先读
README.md— 了解产品定位、五个协议字段、三条隐私红线、安装与工具(公开版 README)。 - 再读
PROGRESS.md— 了解已完成到哪一步、关键决策与理由、下一步、待用户确认事项。 - 禁止重做已完成的工作。 M1–M4 已全部完成,v0.1 已公开发布(见下方发布状态、PROGRESS.md 验收证据)。除非用户明确要求返工,不要重建已完成的里程碑。
发布状态(v0.1 已上线 + 上架官方 Registry · 2026-07-15)
- GitHub: https://github.com/gzchenhao/openhire (owner
gzchenhao,main,tag v0.1.1) - Release v0.1.0: https://github.com/gzchenhao/openhire/releases/tag/v0.1.0 (含快照资产
openhire-index.db.gz,URL 稳定不变) - PyPI: https://pypi.org/project/openhire/0.1.1/ (
pipx install openhire) - 官方 MCP Registry:
io.github.gzchenhao/openhirev0.1.1(registry.modelcontextprotocol.io,用server.json+mcp-publisher发布;PulseMCP/mcp.so 自动同步)。 - Smithery: v0.1 放弃(无本地 stdio 网页入口,见
reports/010)。 - 推送用
gh(keyring);PyPI token 仅%USERPROFILE%\.pypirc;mcp-publisher用其本地 GitHub 凭据。三者均不进代码/git。 - 再发新版流程:改 README
mcp-name保持不变 → bump 版本 →twine upload→ 改server.json版本 →mcp-publisher publish。
常设工作制度(持续遵守)
- 进度记录: 每完成一个里程碑、或每个工作日结束时,更新
PROGRESS.md:- 日期
- 已完成(含验收证据:测试数、实测输出、花费等可核对的事实)
- 关键决策及理由
- 下一步
- 待用户确认事项 控制在一页以内;新进展追加在顶部(倒序),旧条目保留。
- 恢复指引常驻: 本文件顶部的「新会话恢复指引」始终保留并保持最新。
- 每次会话结束前主动提醒用户: 「今日请备份
C:\openhire到 U 盘。」 - 里程碑节奏: 每完成一个里程碑停下,对照验收标准向用户汇报后再继续。
- 汇报归档: 每次完成任务后的完整汇报,除在终端显示外,同时写入
C:\openhire\reports\。- 文件名 = 递增编号 + 主题,如
001-真机验收三件套.md、002-M4打包.md。 - Markdown 格式,自足完整:不依赖终端上下文,单独打开也能看懂(含背景、做了什么、证据、结论、下一步)。
- 每次干完活的最后一行告诉用户:「汇报已写入 C:\openhire\reports\xxx.md」。
- 编号取
reports\里现有最大编号 +1(补零三位)。
- 文件名 = 递增编号 + 主题,如
- 每周快照刷新:已自动化(017)。
.github/workflows/refresh-snapshot.yml每周一 06:10 UTC 自动跑(也可在 Actions 手动 Run workflow):下载已发布快照 → 重新 seed → 免费启发式全量刷新 →ohp snapshot-build→ 覆盖上传同名资产。工作流零密钥(只用 GitHub 自动下发的 per-run token,权限仅contents: write),失败靠 GitHub 默认邮件通知仓库主。 人工只剩月度一条命令的 LLM 精抽(CI 无 key,只能跑启发式):
跑完按ohp extract-rebuild --backend glm # 补 skills(套餐内,现金 ¥0) ohp extract-role-family --backend glm # 补 role_familydocs/maintainer-snapshot-refresh.md手动 build + upload 一次,让精抽结果进到公开快照。
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 · +2 lines · +201 tokens per session 0b58b1bdf961
- 4d ago First seen · 64 lines · 1,743 tokens per session scan A 4e0eda6df992
openhire CLAUDE.md is an instructions file published in the GitHub repository gzchenhao/openhire (2 stars, last pushed yesterday), licensed MIT. It adds 1,944 tokens to every session, about $0.0097 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 instructions, from other repositories
jobseek AGENTS.md
AGENTS.md instructions for colophon-group/jobseek, covering agents.md — jobseek, project overview, repository structure, commands and ops routines (codex-first, claude-compatible).
please-hire-me CLAUDE.md
Instructions for alecswang/please-hire-me, covering please-hire-me — agent entry point (read this first), setup, step by step, rules during onboarding, what this is and repo layout.
ai-job-agent CLAUDE.md
Instructions for AkbarDevop/ai-job-agent, covering ai job application agent - claude code instructions, bundled skills, unified mode — npm run agent, npm scripts and output locations (all gitignored).
jobseek copilot-instructions.md
Copilot instructions for colophon-group/jobseek, a project described as: Monitors company career pages for new job postings and surfaces them in a single dashboard.
jobseek CLAUDE.md
Claude Code instructions for colophon-group/jobseek, a project described as: Monitors company career pages for new job postings and surfaces them in a single dashboard.
awesome-copilot-id AGENTS.md
AGENTS.md instructions for GulajavaMinistudio/awesome-copilot-id, covering communication, explanation and documentation, markdown formatting, user communication style and workflow & methodology.