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/songyang0603/ds-spec-loop/agents-mdgit clone --depth 1 https://github.com/songyang0603/ds-spec-loopWhat 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 | $0.00281 | $0.00281 |
| Opus 5 | $0.00140 | $0.00140 |
| Sonnet 5 | $0.00056 | $0.00056 |
| Haiku 4.5 | $0.00028 | $0.00028 |
Grade A, and why
ds-spec-loop 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 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.
What it actually says
Repository instructions
This repository publishes one portable, repository-native Spec programming Skill.
Authority
skills/ds-spec-loop/SKILL.mdowns the agent-facing core loop.skills/ds-spec-loop/references/owns detailed protocols and templates.README.mdandREADME.zh-CN.mdown public installation and usage guidance and remain equivalent in meaning.docs/decisions/owns this repository's working proposals and accepted decision rationale.NOTICEowns method attribution and independence from upstream projects.
Working rules
- Every non-mechanical change adds or updates one owning decision record in the same bounded change. Search for an existing owner before creating another.
- Reuse the repository's current terminology and tools. Do not create a second Spec, ADR, documentation, test, or CI system when an existing one owns the responsibility.
- A working proposal may change before delivery. A stable accepted decision is replaced by a new cross-linked decision record rather than rewritten as though it never existed.
- Supersession is a relationship, not a lifecycle status or required directory.
- Keep one canonical home per fact. Other files may summarize or link to the owner.
- Use the smallest direct checks that cover the changed claims. Development-only evaluation artifacts stay outside the repository.
- Do not modify
../deepseek-harnesswhile maintaining this project.
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 · 22 lines · 281 tokens per session scan A 6cfc66e9489f
ds-spec-loop AGENTS.md is an instructions file published in the GitHub repository songyang0603/ds-spec-loop (38 stars, last pushed 16d ago), licensed MIT. It adds 281 tokens to every session, about $0.0014 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
desktop-cc-gui AGENTS.md
AGENTS.md instructions for zhukunpenglinyutong/desktop-cc-gui, covering 项目规则入口（mossx）, 规则优先级, 文档分层, 最小读取路径 and openspec 交付.
project-blueprint AGENTS.md
Instructions for shuguang1994/project-blueprint, covering project blueprint 项目开发规范 (agents.md), 一、项目身份, 二、常用命令, 本仓库无编译/测试命令，常用操作为 git 双远程推送： and 技能安装/更新（验证对外文档描述一致性时参考）：.
project-os-for-codex AGENTS.md
Instructions for herry2059/project-os-for-codex, covering ai contributor rules, product truth, architecture and security, development workflow and agent integration contract.
AmbyKit CLAUDE.md
Instructions for ambystechcom/AmbyKit, covering claude code — notes for the ambykit repo and claude-specific.
summono CLAUDE.md
Instructions for liustack/summono, a project described as: One click and DeepSeek Harness is running — the free launcher & installer for AI harnesses. 一键安装启动 DeepSeek Harness。.
ccteam AGENTS.md
Instructions for firstintent/ccteam, covering agents.md — ccteam 实现导引, 〇、架构总览(定向图;红线唯一清单 = §三), 一、需求与版本管理(两条来源,都住 gitignored docs-local/), 二、仓库内文档(只有用户面;架构与协议一律读代码) and 三、不可触碰的架构红线(唯一权威清单).