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/alessandrofare/fable-frame/agents-mdgit clone --depth 1 https://github.com/AlessandroFare/fable-frameWhat 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 | $0.01338 | $0.01338 |
| Opus 5 | $0.00669 | $0.00669 |
| Sonnet 5 | $0.00268 | $0.00268 |
| Haiku 4.5 | $0.00134 | $0.00134 |
Grade A, and why
fable-frame 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.
How it starts
The opening of the file, as written. The whole thing — 75 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md
Guidance for AI coding agents (Claude Code, Minimax Code, OpenCode, Codex, Cursor, Windsurf) working in this repository or consuming this skill.
What this repo is
A universal AI agent skill implemented entirely as Markdown. The runtime artifact is SKILL.md: the agent reads its YAML frontmatter (metadata, license, allowed tools) followed by the three-section behavior + workflow + templates prompt. There is no build step and no code to run.
The skill is designed to be pasted as system instructions, custom instructions, or a project's CLAUDE.md / AGENTS.md. The whole product lives in SKILL.md.
Key files
SKILL.md: the skill itself. YAML frontmatter followed by three sections (Fable Brain behavior, Karpathy principles workflow, 50 templates). This is the source of truth.README.md: for humans. Installation per agent, usage, the full template table, and version history.LICENSE: MIT.
When to activate each section
The agent should treat each section as a different load class:
- Section 1, Core Behavior (Fable Brain). Always on. No trigger needed. These 8 rules apply to every response, in every conversation, regardless of what the user asked for. This is the agent's base temperament.
- Section 2, Karpathy Workflow. Activate for any task that is multi-step, agentic, has an objective metric, involves running multiple sub-tasks, or includes a
program.mdblock. If the task is a single quick question or a one-shot edit, this section is optional context. - Section 3, Templates. Activate only when the user invokes a template with
use template [NAME]. Outside of that trigger, the templates are dormant reference material.
Template invocation flow
When the user types use template [NAME]:
- Load the matching template block from Section 3.
- Identify every
[placeholder]in the template. If the user already filled some in their message, keep those and don't re-ask. - Ask for the missing placeholders in a single message. Group related ones, but do not split across multiple turns. Do not start executing yet.
- Once placeholders are provided, execute the template exactly as written, in the order it specifies. Treat the template's "Deliver" list as the contract.
- If a deliverable is ambiguous, ask one focused clarifying question. If the template's role prompt (e.g. "You are a senior backend engineer") conflicts with Section 1 behavior rules, Section 1 wins: the agent still grounds claims, leads with the outcome, and stops at real boundaries.
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 · 75 lines · 1,338 tokens per session scan A 3dccf6a0700b
fable-frame AGENTS.md is an instructions file published in the GitHub repository AlessandroFare/fable-frame (6 stars, last pushed 2mo ago), licensed MIT. It adds 1,338 tokens to every session, about $0.0067 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
agentic-ai-brain-for-trending AGENTS.md
AGENTS.md instructions for tuanduongdn1504/agentic-ai-brain-for-trending, covering storm bear — agent instructions, quick redirect, why this shim exists, context parity guidelines (startup rules) and vault state architecture.
ii-agent AGENTS.md
Instructions for Intelligent-Internet/ii-agent, covering ii-agent contributor guide, quick start, repository map, mandatory rules and architecture.
intelligent-terminal rust.instructions.md
Concise Rust coding conventions for this repository.
Browser4 CLAUDE.md
Instructions for platonai/Browser4, covering browser4 — project context for claude, architecture, key dispatch chain (cli → browser), batch commands and e2e test structure.
nimbalyst AGENTS.md
AGENTS.md instructions for nimbalyst/nimbalyst: At the beginning of every AI session, read the root CLAUDE.md and any applicable package-level CLAUDE.md before taking actions.
sortie copilot-instructions.md
Instructions for sortie-ai/sortie, covering sortie coding & review standards, 1. layered imports (downward only; violation is critical), 2. concurrency safety, 3. workspace path safety (critical: security boundary) and 4. persistence (sqlite).