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/injaytseng/graph-engineering-on-research/agents-mdgit clone --depth 1 https://github.com/InjayTseng/graph-engineering-on-researchWrote 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/injaytseng/graph-engineering-on-research/agents-md)<a href="https://agentmods.dev/instructions/injaytseng/graph-engineering-on-research/agents-md"><img src="https://agentmods.dev/badge/instructions/injaytseng/graph-engineering-on-research/agents-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 | $0.01308 | $0.01308 |
| Opus 5 | $0.00654 | $0.00654 |
| Sonnet 5 | $0.00262 | $0.00262 |
| Haiku 4.5 | $0.00131 | $0.00131 |
Grade A, and why
graph-engineering-on-research 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 today.
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 — 82 lines — stays where its author put it; the contents beside it link to each section on GitHub.
graph-engineering-on-research — agent bootstrap
Public MIT repo：七份可直接複製貼上的 prompt 範本，把 AI agent 從排隊變成平行開火的圖， 再把圖反過來攻擊自己的結論。同時打包成 Claude Code plugin，每份範本變成一個 slash command。
任何 AI session 或外部貢獻者動這個 repo 前先讀完本檔。
這是公開 repo，所以本檔刻意不放 maintainer 的內部路徑、任務系統或排程——那不是疏漏， 是邊界。在 maintainer 機器上工作的 session 另有一份 vault 側的契約要遵守，以那份為準； 不要為了「補完」而把任何內部路徑貼進這個檔案。
1. 身分
| 項目 | 值 |
|---|---|
| Canonical remote | github.com/InjayTseng/graph-engineering-on-research（public, MIT） |
某些 clone 的資料夾名是 graph-engineering-sample，與 repo 名不同。以
git remote get-url origin 認身分，不要用資料夾名猜——這是本 repo 最常見的接手錯誤。
2. 狀態（2026-09-01 實查）
- 開發在
main上，變更走 PR 進來（近期每個 commit 都帶(#NN)）。 python3 scripts/check-invariants.py9 checks pass。- 接手前自己跑一次
git status -sb——本節不記錄 HEAD，記了隔一個 commit 就是錯的。
3. 這個 repo 的規則是可執行的，不是散文
.github/workflows/ci.yml 把 README 的宣稱與 docs/adr/ 的決策變成 merge 前的閘門。
scripts/check-invariants.py 目前跑九項：
- ADR 0001 — 每個 copy block 的使用者輸入都放在最後
- ADR 0004 — skill 不自我呼叫
- prompts / skills / 兩個語言版本保持同步
.graph/ledger.md有被讀也有被追加docs/diagrams/每個來源都畫了、每張圖都被嵌入docs/demo/的 GIF 有嵌入且在容量預算內.graph/ledger.md存在、四欄位、只有條目examples/每份逐字稿都在兩份 README 裡被連到- 兩個語言版本的 README 成對
另有兩道不在那九項裡：
- 每次 push 與 PR 都跑：重建圖後
git diff --exit-code。build.py的輸出是 byte-identical 的，所以只要 commit 的圖跟重建結果有一 byte 不同就擋下。 - 只在 PR 上跑：
.graph/ledger.md是 append-only，CI 用 byte prefix 比對確認舊行沒被 改過或截斷。判斷錯了就加一行新的日期記錄，原本那行留著。
改任何 prompt、README 或圖之前先在本機跑一次 invariants，比等 CI 快。
4. 跑起來與驗證
python3 scripts/check-invariants.py # 9 checks（2026-09-01 實跑 pass）
python3 docs/diagrams/build.py # 重建 SVG + HTML
python3 docs/diagrams/build.py --png # 加上 2x 點陣化，需要 playwright
圖是產生出來的，不是手畫的。在乾淨的 main 上重跑 build.py 後 git status 應該是空的
（輸出是 byte-identical 的）。如果不是，代表來源或環境有變，先查清楚再 commit。
5. 語言與結構紀律
README.md（英文）是預設入口；README.zh-TW.md服務中文圈。兩份必須成對更新，CI 會擋。prompts/七份範本（00–06）、skills/七個對應的 plugin skill、.claude-plugin/是 plugin 清單。三者必須同步，CI 會擋。examples/是實跑逐字稿，含出錯的部分。那是刻意保留的，不要「整理乾淨」。docs/adr/六份決策記錄。要違反其中任何一條，先改 ADR 並說明理由，不要直接繞過。
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.
- today First seen · 82 lines · 1,308 tokens per session scan A ac01b363d305
graph-engineering-on-research AGENTS.md is an instructions file published in the GitHub repository InjayTseng/graph-engineering-on-research (10 stars, last pushed 4d ago), licensed MIT. It adds 1,308 tokens to every session, about $0.0065 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 2026-09-04.
Other instructions, from other repositories
open-multi-agent AGENTS.md
AGENTS.md instructions for open-multi-agent/open-multi-agent, covering agents.md, repository map, commands, working rules and validation by change type.
commonly CLAUDE.md
Claude Code instructions for Team-Commonly/commonly, covering claude.md / agents.md, 🧠 product vision & architecture philosophy, what commonly is, the architecture model and key concepts.
research-os AGENTS.md
Instructions for lxinfei5/research-os, covering researchos — constitution, §0 what this is, §1 directory discipline, §2 half-life (innovation 1) — memory design and §3 behavior pillars (innovations 2–5).
opencode-hive AGENTS.md
Instructions for rretsiem/opencode-hive, covering opencode hive, directory structure, ai setup instructions, step 1: discover available models and step 2: detect target project stack.
agent-flow AGENTS.md
AGENTS.md instructions for josix/agent-flow, covering codex co-review — repo guide, project context, your role: phase 4 co-reviewer, output contract and severity scale.
prism-insight CLAUDE.md
Claude Code instructions for dragon1086/prism-insight, covering claude.md - ai assistant guide for prism-insight, quick overview, project structure, analysis pipeline and ai agents.