</>MCP Agents Market
Plugin

fablize

by fivetaku895PythonUpdated 2026-07-06

A Claude Code plugin that makes Opus behave like Fable — completion, evidence, and verification enforced as procedure. Ships only what a Fable-vs-Opus comparison proved transferable.

Claude Code

Fablize is a Claude Code plugin that enforces completion, evidence gathering, and verification procedures in Claude Opus and other Claude models. Born from a controlled comparison between Fable 5 and Opus 4.8 across 1,500+ tool calls, it transfers only the procedural behaviors proven to be effective—verification grounding, multi-story decomposition, systematic investigation, and early-stop prevention. The plugin does not claim to enhance model capabilities but ensures Claude reaches its existing ceiling by running artifacts, demanding evidence before marking tasks complete, and escalating when capability limits are reached.

Key Features

Verification grounding that runs and observes render/executable artifacts (HTML, SVG, games, charts) before completion
Multi-story decomposition with an evidence gate that refuses completion without proof of progress
Systematic investigation protocol: reproduce bugs, generate competing hypotheses, trace full causal chains
Early-stop hook that blocks "promising-without-doing" behavior when Claude announces intent but doesn't execute
Per-task router that injects only the matching verified discipline based on work type (debugging, rendering, multi-step tasks)
Trigger modes: explicit via /fablize command, phrase detection ("see it through"), or always-on after setup
Transparent capability ceiling detection with escalation prompts to stronger models or human review
Built on 45-run controlled experiment data, shipping only procedures with measured effectiveness

Use Cases

  • 01Building interactive HTML/SVG visualizations where you need Claude to verify rendering before claiming completion
  • 02Multi-step code projects requiring decomposition into verifiable checkpoints with evidence at each stage
  • 03Debugging sessions that need systematic root-cause analysis rather than guess-and-patch attempts
  • 04Game or chart development workflows where execution validation is critical before handoff
  • 05Complex coding tasks where you want to prevent premature "done" declarations without proof
  • 06Projects at the edge of model capability where you need explicit escalation signals instead of hallucinated solutions

Related Plugins

View more

andrej-karpathy-skills

A single CLAUDE.md file to improve Claude Code behavior, derived from Andrej Karpathy's observations on LLM coding pitfalls.

Plugin216.6k

claude-mem

Persistent Context Across Sessions for Every Agent – Captures everything your agent does during sessions, compresses it with AI, and injects relevant context back into future sessions. Works with Claude Code, OpenClaw, Codex, Gemini, Hermes, Copilot, OpenCode + More

Plugin95.2k

Understand-Anything

Graphs that teach > graphs that impress. Turn any code into an interactive knowledge graph you can explore, search, and ask questions about. Works with Claude Code, Codex, Cursor, Copilot, Gemini CLI, and more.

Plugin85k

rtk

CLI proxy that reduces LLM token consumption by 60-90% on common dev commands. Single Rust binary, zero dependencies

Plugin82.2k

fablize — FAQ

What is fablize and what does it do?+

Fablize is a Claude Code plugin that makes Claude Opus (and other Claude models) follow verification and completion procedures derived from Fable 5's working patterns. It enforces running artifacts, demanding evidence before completion, and systematic investigation—transferring procedural discipline without claiming to enhance core model intelligence.

How do I install the fablize plugin in Claude Code?+

Use the commands /plugin marketplace add fivetaku/fablize then /plugin install fablize in Claude Code. For always-on operation, run the setup script at ${CLAUDE_PLUGIN_ROOT}/setup/setup.sh and choose local or global mode.

Which AI clients work with fablize?+

Fablize is designed specifically for Claude Code as a plugin. It hooks into Claude's task execution flow through UserPromptSubmit events and per-task routing.

Does fablize require API keys or prerequisites?+

No external API keys are needed. You only need Claude Code with plugin support installed; fablize registers its hooks automatically upon installation.

Is fablize free to use?+

Yes, fablize is open source and released under the MIT license, free for any use including commercial projects.

Can fablize make Claude as smart as Fable on open-ended discovery?+

No. The README explicitly states that out-of-spec defect discovery, open-ended creative detail, and self-driven propagation depth are model capabilities that cannot be transferred by a harness. Fablize transfers only procedural discipline—execution, verification, investigation—that Opus was capable of but inconsistently applied.

How do I install fablize?+

Open the source repository on GitHub and follow its README. fablize is a plugin — MCP Agents Market links you directly to the official repo.

Is fablize free?+

fablize is an open-source project hosted on GitHub. Check the repository for its license and any usage requirements.

Related searches

Developer Tools PluginsTesting PluginsProductivity & Workflow PluginsBest MCP ServersBest SkillsBest AgentsTop rated listingsClaude Code toolsBrowse all categories