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/jmet04/ai-apprentice/agents-mdgit clone --depth 1 https://github.com/JMET04/ai-apprenticeWhat 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 | $0.14761 | $0.14761 |
| Opus 5 | $0.07380 | $0.07380 |
| Sonnet 5 | $0.02952 | $0.02952 |
| Haiku 4.5 | $0.01476 | $0.01476 |
Grade A, and why
ai-apprentice 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 2d 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 — 251 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md
Project Vision
This project is called Transparent AI Apprentice MCP.
The goal is to build a universal teachable AI apprentice system. The product should allow humans to teach AI like teaching a student or training a new assistant. The AI should learn from visual demonstrations, natural language corrections, structured feedback, examples, and execution history.
The goal also includes a knowledge-augmented RAG research lane: ordinary users should be able to low-cost cultivate domain-specialized apprentices by providing manuals, standards, papers, software docs, teacher notes, prior corrections, examples, and log-format references. RAG is evidence only: it may retrieve source-backed context and draft disabled rules, but it must not enable rules, execute software, bypass teacher confirmation, write long-term memory, claim technology acceptance, or unlock packaging.
The core product principle is:
Human teaches -> AI executes -> human corrects -> system extracts rules -> AI improves -> execution becomes more controllable and transparent.
Product Principles
- Do not design this as a generic chatbot.
- Do not design this as a cold workflow automation tool.
- Design it as an AI apprentice that can be taught, corrected, and trusted.
- Every important AI action should produce a visible, structured trace.
- Never expose private chain-of-thought. Show structured reasoning traces instead: steps, rules, confidence, validation results, and human review points.
- Human feedback should become reusable memory.
- The user should always understand why the AI produced a result.
- Retrieved knowledge should become provenance-carrying evidence packets, not hidden prompt magic or automatic authority.
Architecture Principles
Keep these layers separate:
- UI layer
- API layer
- AI service layer
- workflow execution engine
- memory store
- trace store
- correction/rule extraction module
- knowledge source registry and retrieval evidence packet layer
- tool registry
- skill registry
- guardrail/policy layer
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.
- 2d ago First seen · 251 lines · 14,761 tokens per session scan A 3f4be228b41a
ai-apprentice AGENTS.md is an instructions file published in the GitHub repository JMET04/ai-apprentice (1 stars, last pushed 12d ago), licensed MIT. It adds 14,761 tokens to every session, about $0.0738 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
vscode buildNext.instructions.md
Working notes and architecture documentation for the new esbuild-based build system in build/next. Use when making changes to the new build pipeline (transpile/bundle commands, NLS plugin, source-map handling, resource copying, or self-hosting watch tasks).
spec-kit AGENTS.md
AGENTS.md instructions for github/spec-kit, covering agents.md, about spec kit and specify, quickstart — add a new integration in 5 steps, integration architecture and integrationmanifest — file tracking.
codex AGENTS.md
AGENTS.md instructions for openai/codex, covering rust/codex-rs, the codex-core crate, code review rules, crate api surface and model visible context.
langchain AGENTS.md
AGENTS.md instructions for langchain-ai/langchain, covering global development guidelines for the langchain monorepo, corridor security analysis, project architecture and context, monorepo structure and development tools & commands.
vscode oss-third-party-notices.instructions.md
Instructions for microsoft/vscode, covering vs code oss third-party-notices pipeline, architecture, pipeline flow in ci, applying the notice (cutover) and fallback chain (never fail the build).
next.js AGENTS.md
Instructions for vercel/next.js, covering next.js development guide, codebase structure, monorepo overview, core package: packages/next and other important packages.