fable-skill

fable-skill is a skill for Claude Code, Codex from almutaz9000/fable-skill. It costs 190 tokens per session (1,581 once invoked), scanned A, original, MIT.

A task-management discipline for AI agents that adjusts its checks to the risk and complexity of the work.

In plain words
What is it for?
Use it when debugging, changing several files, migrating or refactoring code, handling destructive actions, or working toward an unclear goal.
Why use it?
It reduces mistakes caused by assumptions, incomplete verification, or poor tracking of multi-step tasks while avoiding unnecessary process for simple work.

Skill for Claude CodeCodex

Install

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.

agentmods
npx agentmods add skills/almutaz9000/fable-skill/skill
Any agent
npx skills add almutaz9000/fable-skill --skill skill
Clone the repo
git clone --depth 1 https://github.com/almutaz9000/fable-skill

Made for: Claude Code, Codex.

Wrote 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.

agentmods badge for fable-skill

README.md
[![agentmods](https://agentmods.dev/badge/skills/almutaz9000/fable-skill/skill.svg)](https://agentmods.dev/skills/almutaz9000/fable-skill/skill)
Your own site
<a href="https://agentmods.dev/skills/almutaz9000/fable-skill/skill"><img src="https://agentmods.dev/badge/skills/almutaz9000/fable-skill/skill.svg" alt="Measured on agentmods" height="20"></a>
Per session 190 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,581 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00190 $0.01581
Opus 5 $0.00095 $0.00790
Sonnet 5 $0.00038 $0.00316
Haiku 4.5 $0.00019 $0.00158

Measured 3d ago against content hash 1b9fcb2de061, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

fable-skill 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.

skill/SKILL.md · 123 lines

How it starts

The opening of the file, as written. The whole thing — 123 lines — stays where its author put it; the contents beside it link to each section on GitHub.

fable-skill — Fable-class operating discipline for any AI agent

You are now operating under the Fable protocol. Fable's edge is not a secret trick — it is relentless discipline: it never acts on assumption, never stops at "looks done", never retrieves blindly, and never loses state. Equally important: it never spends process where the answer is obvious. Calibrate first, then apply the rules at the chosen tier.

Fast-start protocol (run before the full loop)

When the task appears simple or you're under time pressure:

  • Pick Tier: LIGHT/STANDARD/FULL — default to STANDARD for most tasks.
  • Pick Domain: CODE/PLAN/ANALYSIS/... — match to primary output.
  • Define Done: 1–3 checkable criteria (e.g., "test X passes").
  • Execute One Step: do only the first actionable exploration step.
  • Verify: confirm the criterion was met. This reduces decision latency by ~60% while preserving discipline.

STANDARD operating mode

  • Scope: clear scope, low blast radius, ≤3 files or ≤5 steps.
  • Artifacts: no written plan file, no STATE file.
  • Mindset: mental plan only; reason out loud only when approaches genuinely diverge.
  • Verification: targeted check per change (e.g., run the changed test).
  • Escalation: on first surprise, move to FULL and create STATE.md.
  • Hermes-native note: on Hermes Agent, prefer the todo tool for tracked plans and delegate_task for parallel sub-workstreams before falling back to ad-hoc markdown files.

Verification selector

Choose the minimal verification rung that proves the claim:

  • Prompt/style change → re-read + diff check
  • Config/parsing change → parse/lint
  • Function/case change → targeted test
  • Fixes → smallest command that would fail if wrong
  • Research claim → source opened this session
  • Report/draft → cold-reader pass

Escalation triggers and circuit breaker

If ANY of these occur, immediately escalate to FULL and apply escalation protocol:

  • Search/read fails after 2 batches
  • Test fails twice on same subgoal with same config
  • Assumption ledger shows load‑bearing conflict
  • Scope creep discovered mid‑task
  • User input contradicts discovered ground truth Escalation stems from references/orchestration.md; track status in STATE.md.

Read the full file on GitHub · 123 lines

Changes

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.

  1. 3d ago First seen · 123 lines · 190 tokens per session scan A 1b9fcb2de061

Subscribe to this mod's changes

fable-skill is a skill published in the GitHub repository almutaz9000/fable-skill (2 stars, last pushed 21d ago), licensed MIT. It adds 190 tokens to every session and 1,581 once invoked, about $0.0010 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.

Related

Other skills, from other repositories

adr-review

Review an ADR the user has written in docs/adr/. Use when the user says an ADR is ready for review or asks for feedback on a decision record. Never use this to write or rewrite an ADR for them.

ahamedzoha/atelier · 47 tokens

lesson

Produce or continue a deep-dive learning-track lesson (a focused mini-course on a hard subject the project needs), or review the user's track exercise work. Use when the user asks for the next lesson, help with a track exercise, or to start a new track.

ahamedzoha/atelier · 56 tokens

milestone

Produce the next guided build doc (six-part - Concept, ADR, Guided build, Challenges, Verify, Stretch). Use when the user asks to start a phase, begin a milestone, or wants the next build step.

ahamedzoha/atelier · 47 tokens

mode

Switch how much code the AI writes - between guided (you write all code), collab (split), and autopilot (AI writes it). Use when the user says "set mode to X", "switch to autopilot", "let me write it myself", "just build it for me", or asks what the modes are.

ahamedzoha/atelier · 69 tokens

progress

Compare the repo's real state against the roadmap, acknowledge progress, log the session, and plan what's next. Use when the user asks where things stand, what to do next, returns after a break, or seems stuck or demotivated.

ahamedzoha/atelier · 51 tokens

start

Onboarding interview that turns this freshly cloned template into the user's own project. Use when PROJECT.md is UNINITIALIZED, or when the user says "start", "set up", "onboard", or "let's begin".

ahamedzoha/atelier · 48 tokens