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/jiadizhunine/deeptutor/agents-mdgit clone --depth 1 https://github.com/jiadizhunine/deeptutorWrote 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/jiadizhunine/deeptutor/agents-md)<a href="https://agentmods.dev/instructions/jiadizhunine/deeptutor/agents-md"><img src="https://agentmods.dev/badge/instructions/jiadizhunine/deeptutor/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 5.1 | $0.03612 | $0.03612 |
| Opus 5 | $0.01806 | $0.01806 |
| Sonnet 5 | $0.00722 | $0.00722 |
| Haiku 4.5 | $0.00361 | $0.00361 |
Grade A, and why
deeptutor 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 6d 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 — 244 lines — stays where its author put it; the contents beside it link to each section on GitHub.
DeepTutor — Academic Advisor Investigation System
Single source of truth for agents.md compliant tools (Codex CLI, OpenCode, OpenClaw, Aider, Cline, Continue, etc.). For Claude Code and Cursor, see
SKILL.mdand.cursor/rules/deeptutor.mdc— both reference this file's workflow.
Comprehensive professor evaluation tool for graduate-school decisions. Investigates publications, co-author networks, student trajectories (the single most predictive signal), lab culture, and exploitation/PUA risk. Produces a standalone Chinese or English .html report.
Core principle: Your ceiling = your seniors' ceiling. A professor whose students thrive is gold; one whose students vanish is a red flag — regardless of publication metrics.
What this repo is
This is not a traditional code project — it is an AI "skill" / agent prompt-pack. There is nothing to build or npm install. When a user asks the assistant to evaluate a professor, the assistant should:
- Detect language and region from the user's input.
- Run the 10-phase investigation workflow below (or the 6-phase Lite variant for lower-capability models).
- Output structured JSON, then render to HTML via
scripts/generate_report.py.
The Python scripts in scripts/ are self-contained helpers — their enabled network path uses only Python's standard library. The fetcher pins each validated hostname to the public IPs returned by one DNS lookup, and search_social.py must reuse that fetcher rather than opening URLs independently. Its curl_cffi and Playwright interfaces fail closed because this integration cannot prove that every optional network path preserves the same boundary. The bundled fetcher intentionally refuses local/private/link-local targets.
When to trigger
Activate this workflow whenever a user asks to investigate, evaluate, or compare graduate advisors. Trigger phrases include:
- Chinese:
调查导师,评估教授,选导师,导师怎么样,能不能跟这个老师读研,这个导师push吗,帮我看看XX老师,对比这几个导师 - English:
investigate this advisor,should I join this lab,evaluate professor,is this prof good,rate my potential advisor,review this PI,compare these advisors - Implicit: user provides
<professor name> + <institution>with any evaluation intent.
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.
- 6d ago First seen · 244 lines · 3,612 tokens per session scan A 95a061eefd54
deeptutor AGENTS.md is an instructions file published in the GitHub repository jiadizhunine/deeptutor (29 stars, last pushed 1mo ago), licensed MIT. It adds 3,612 tokens to every session, about $0.0181 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
AlterLab-Academic-Skills CLAUDE.md
Claude Code instructions for AlterLab-IEU/AlterLab-Academic-Skills, covering project overview, audience, naming convention, skill categories and core pipeline routing rules.
phdtaketaketake AGENTS.md
Instructions for powerofjinbo/phdtaketaketake, covering phdtaketaketake — agent instructions, use, hard rules (non-negotiable for both workflows) and installed clis (after pip install -e .).
degreeworks-cli AGENTS.md
Instructions for Aaryan-Kapoor/degreeworks-cli, covering agents.md — degreeworks cli, tl;dr, schedule planning protocol, phase 0 — verify auth and phase 1 — gather ground truth.
degreeworks-cli CLAUDE.md
Instructions for Aaryan-Kapoor/degreeworks-cli, a project described as: Read-only DegreeWorks CLI that lets an AI agent build your exact KSU semester schedule — real sections, times, CRNs, verified prereqs — from your degree audit.
open-dictionary AGENTS.md
Instructions for ahpxex/open-dictionary, covering open dictionary rewrite charter, product framing, core workflow, technical framework and 1. raw ingestion layer.
ab900 CLAUDE.md
Instructions for timothywarner-org/ab900, covering claude.md, repository purpose, architecture, commands and 2. non-ascii punctuation -- must return zero.