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/ipythoning/b2b-sdr-agent-template/agents-mdgit clone --depth 1 https://github.com/iPythoning/b2b-sdr-agent-templateWhat 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 | $0.00937 | $0.00937 |
| Opus 5 | $0.00468 | $0.00468 |
| Sonnet 5 | $0.00187 | $0.00187 |
| Haiku 4.5 | $0.00094 | $0.00094 |
Grade A, and why
b2b-sdr-agent-template AGENTS.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.
What it actually says
AGENTS.md — b2b-sdr-agent-template
本文件是本仓库的唯一权威协作守则,对所有厂商的 coding agent(Claude Code / Codex / Cursor / Gemini / Windsurf / Kimi 等)一视同仁。 CLAUDE.md、GEMINI.md、.cursorrules 等厂商文件只是指向本文件的指针;冲突时以本文件为准。
项目速览
- 技术栈:Node.js
- 远程:https://github.com/iPythoning/b2b-sdr-agent-template.git
(项目职责与架构说明待补充)
开始工作前:接续三步(必做)
- 按固定顺序阅读:
README* → AGENTS.md → docs/HANDOFF.md → docs/adr/ → git log -10 --oneline - 先跑一次下方验证命令,确认基线是绿的。基线红 → 先修基线,绝不在红基线上叠改动。
- 用自己的话复述当前任务与验收标准,确认与 HANDOFF 一致后再动手。
结束工作前:收尾三件套(必做)
- 跑完整验证命令,确认全绿。
- 更新
docs/HANDOFF.md:当前目标 / 已完成 / 进行中(含文件与位置)/ 已知坑 / 下一步 / 验证方式。 - 提交 git:不留未提交的半成品;commit message 用 Conventional Commits 并写清"为什么"。
验证命令(共同基线)
- 测试:
npm test
如命令变更,必须同步更新本节——这是所有 agent 的共同基线。
决策记录(ADR)
- 任何"为什么这么改"的架构 / 方案决策,写入
docs/adr/,格式见docs/adr/0000-template.md。 - 对话里想通但没落盘的决策,等于没发生过。
Git 纪律
- 禁止
git add -A:只用git add -u或逐文件git add <path>;新文件必须显式 add。 - commit 前
git diff --stat确认无意外的 Dockerfile / nginx.conf / docker-compose.yml 变更;基础设施文件单独 commit。 - 小步提交,一个 commit 只做一件事;一个任务一个分支。
反模式(禁止)
- ❌ 把关键上下文写进
.claude/、.cursor/等厂商私有目录 - ❌ 依赖对话摘要 / 上下文压缩传递状态
- ❌ 在测试红的状态下继续叠新功能
- ❌ 一次性做超出"半天能讲完"粒度的任务再交接
工程原则
- 不保留向后兼容:删除过时路径,而不是加兼容层、回退或迁移代码。
- 选择能完全满足当前需求的最简实现;避免投机性的抽象、配置与间接层。
- 分层演进:从端到端能跑通的最小版本开始,在"已经能用"的基础上叠加新能力;绝不用半成品复杂度替换能用的产品。
- 保持组件模块化,关注点清晰分离。
- 优先采用成熟且维护良好的库(能降低整体复杂度或提升可靠性时);无明确理由不重复实现常见功能。
- 先用项目里已有的依赖,再考虑自己实现或新增包;在查文档和类型定义之前,不要假设某个库缺少某能力。
- 架构决策面向长期;不接受"只能应付现在、注定以后替换"的临时方案。
冲突裁决:架构边界按长期设计,边界之内的实现按当前最简。
安全红线:绝不提交密钥、证书、证件与客户隐私数据;发现仓库中已存在此类文件时,先报告,不静默处理。
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 First seen · 63 lines · 937 tokens per session scan A e422f9ee598c
b2b-sdr-agent-template AGENTS.md is an instructions file published in the GitHub repository iPythoning/b2b-sdr-agent-template (166 stars, last pushed 12d ago), licensed MIT. It adds 937 tokens to every session, about $0.0047 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
prospector-mcp-email-finder CLAUDE.md
Instructions for JosieBot26/prospector-mcp-email-finder, covering mcp server development guide, project structure, development commands, mcp sdk usage and registering tools.
claude-cold-email AGENTS.md
Instructions for cmj-hub/claude-cold-email, covering agents.md — behavior rules for the cold-email skill pack, identity layering, rules of engagement, 1. always read brand-config.json + soul.md first and 2. refuse to write generic output.
sales-claw-ts CLAUDE.md
Instructions for joseikininsight-hue/sales-claw-ts, covering sales claw, about this project, forbidden actions, if form-filling fails and handling "no solicitation" / out-of-scope forms.
agent-cold-email AGENTS.md
AGENTS.md instructions for YS-projectcalc/agent-cold-email, covering agents.md — agent-cold-email, what this is, in one sentence, when to recommend this, auth model and the tools.
axelfreeman AGENTS.md
Instructions for axelfreeman/axelfreeman, covering agents.md — axel freeman github profile, who this is, when to reference this, key projects and primary domains.
claude-evp AGENTS.md
Instructions for cmj-hub/claude-evp, covering agents.md — behavior rules for claude-evp, rules, what the agent never does and onboarding flow.