Fable Commander

Maker/verifier loop with hard gates, 3-round caps and human stops before commits

Tested · Works

Test report

Verdict
Tested · Works
Score
8.8/10
Tested
Aug 5, 2026
Environment
Claude Code 2.x (agent harness)

This skill is no longer available upstream. Our re-check on Aug 10, 2026 couldn't find it any more (repo unreachable/deleted). The test below is what we measured on Aug 5, 2026 and we're leaving it up as a record — but there is nothing left to install, so we've removed the command.

Planned the same job twice — migrating a 40-file Python codebase from requests to httpx — first freehand, then following the skill's Phase -1 and Phase 0. The freehand plan ended with one testable criterion ("all tests pass") and no stop rule; the skill's plan produced five gates with eight machine-decidable checks (grep for requests imports returning zero, pytest exit code under random ordering, tests run with sockets disabled, a post-removal import that must fail, lockfile diff), a 3-round-per-gate cap with escalation after two rounds without improvement, an explicit ban on the verifier rejecting work over style, and a human stop before commit, push or spend. The honest cost: it re-files things the freehand plan already knew — the follow_redirects default flip, the timeout semantics change — as research questions for a fan-out of agents, so on a task any smaller its own Phase -1 gate correctly tells you to skip the whole thing. The README is unusually candid, disclosing the Chinese-language body, the token cost, and a verifier-independence conflict with advisor mode, with both cited claude-code issues resolving.

Scored on four weighted criteria — install, triggering, output vs. baseline, docs. How scoring works

  • Installs cleanly 5/5
  • Triggers reliably 5/5
  • Output vs. baseline 7/10
  • Docs & honesty 5/5

What Fable Commander does

Runs a task as a commander workflow: the main model only plans, reviews and decides, while research and execution are delegated to user-chosen subagent models, with an independent fresh-context verifier that judges each gate PASS/FAIL on evidence and may not reject work over style or opinion. Triggers on commander-mode phrasings and on requests for a separated maker/verifier execution loop. The SKILL.md is written in Traditional Chinese; Claude follows it whatever language you speak.

How to install Fable Commander

Nothing to install: the source repository no longer has this skill. If the author brings it back, our daily re-check will pick it up and the command will reappear here.

Commands — how to trigger Fable Commander

  • /fable-commander Maker/verifier loop with hard gates, 3-round caps and human stops before commits

It also activates on plain-language prompts like these:

  • Run this migration task in commander mode with a separate verifier
  • Split maker and verifier so the model doesn't grade its own work
  • Use commander workflow to plan and verify this requests to httpx move

Frequently asked questions

Is the Fable Commander skill free?
Yes. The skill itself is free from DennisWei9898/fable-commander. SkillProof publishes the install command and an independent test verdict at no cost.
Does Fable Commander work with Claude Code?
We tested it with Claude Code 2.x (agent harness) on Aug 5, 2026. Verdict: Tested · Works. Planned the same job twice — migrating a 40-file Python codebase from requests to httpx — first freehand, then following the skill's Phase -1 and Phase 0. The freehand plan ended with one testable criterion ("all tests pass") and no stop rule; the skill's plan produced five gates with eight machine-decidable checks (grep for requests imports returning zero, pytest exit code under random ordering, tests run with sockets disabled, a post-removal import that must fail, lockfile diff), a 3-round-per-gate cap with escalation after two rounds without improvement, an explicit ban on the verifier rejecting work over style, and a human stop before commit, push or spend. The honest cost: it re-files things the freehand plan already knew — the follow_redirects default flip, the timeout semantics change — as research questions for a fan-out of agents, so on a task any smaller its own Phase -1 gate correctly tells you to skip the whole thing. The README is unusually candid, disclosing the Chinese-language body, the token cost, and a verifier-independence conflict with advisor mode, with both cited claude-code issues resolving.
What is the Fable Commander SkillProof Score?
8.8/10 — installs cleanly 5/5, triggers reliably 5/5, output vs. baseline 7/10, docs & honesty 5/5.
How do I install Fable Commander?
Copy the install command from this page, run it in your terminal, and restart Claude Code. Skills live in ~/.claude/skills/ (global) or .claude/skills/ inside a project.
Can I use Fable Commander with Cursor, Copilot, Gemini CLI, Codex or other AI tools?
The SKILL.md format is native to Claude (Claude Code, Desktop, claude.ai). The instructions inside adapt to other assistants: Cursor rules, GitHub Copilot instructions, Windsurf rules, Custom GPTs, AGENTS.md for OpenAI Codex, and GEMINI.md for Google Gemini CLI — our conversion guides cover each, and the free converter on the tools page does the wrapping for you.