</>MCP Agents Market
Skill

fable-mode

by mrtooher857Updated 2026-09-02

A Claude skill that activates Fable-style agentic behavior: explicit multi-stage planning, sub-agent delegation, and self-verification.

Claude Code

Fable-mode is a Claude skill that activates disciplined, multi-stage execution for complex tasks requiring decomposition, delegation, and verification. It guides Claude through explicit planning, parallel sub-agent delegation (when runtime tooling supports it), failable verification checks at each stage, and self-critique before delivery. The skill includes four variants (default, Opus, Sonnet, Haiku) that either run inline or spawn model-specific subagents. Developers install it by placing the skill directories into their Claude environment's skill-loading path, where it triggers automatically on multi-file or multi-session tasks.

Key Features

Enforces staged execution with upfront written plans before action
Parallel sub-task delegation to lower-tier models where runtime supports the Agent tool
Failable verification checks at each stage (not sentiment-based validation)
Skeptical self-review and critique before final delivery
Four variants: inline (fable-mode), Opus-pinned, Sonnet-pinned, and Haiku-pinned subagents
Automatic triggering on multi-file, multi-source, or multi-session tasks
Checklist-style structure that imposes discipline without altering model weights
Includes worked before/after example showing verification catching errors

Use Cases

  • 01Large refactoring tasks spanning multiple files or modules
  • 02Multi-source research synthesis requiring independent data gathering
  • 03Long-running development work split across multiple sessions
  • 04Complex analysis tasks where systematic verification prevents errors
  • 05Cost-sensitive high-volume work requiring structured execution (Haiku variant)
  • 06Peak-quality synthesis tasks requiring Opus-level reasoning with Sonnet/Haiku sub-workers

Related Skills

View more

fable-mode — FAQ

What is fable-mode?+

Fable-mode is a Claude skill that enforces structured execution discipline on complex tasks by requiring explicit planning, stage-by-stage verification, sub-agent delegation, and self-critique. It shapes the procedure Claude follows but does not increase the underlying model's reasoning capability.

How do I install fable-mode for Claude?+

Place each skill directory (fable-mode and optionally fable-opus, fable-sonnet, fable-haiku) into the skills directory your Claude environment reads from, such as a skills folder used by Claude Code. The folder name must match the skill's frontmatter name field.

Which Claude clients support fable-mode?+

Fable-mode works with any Claude environment that loads skills, such as Claude Code. The Opus, Sonnet, and Haiku variants require a runtime with the Agent tool to spawn subagents.

Do I need API keys or special prerequisites?+

No API keys are required for the default fable-mode variant. The Opus, Sonnet, and Haiku variants require a Claude runtime that supports the Agent tool for subagent delegation.

Is fable-mode free to use?+

Yes, fable-mode is open source and free to use. Standard Claude API or subscription costs apply based on which model variant you choose and your runtime environment.

When should I use fable-mode versus running Claude normally?+

Use fable-mode for multi-file, multi-source, or multi-session tasks requiring systematic verification. Skip it for simple, single-pass tasks with one obvious approach, as staging adds overhead without benefit for trivial work.

How do I install fable-mode?+

Open the source repository on GitHub and follow its README. fable-mode is a skill — MCP Agents Market links you directly to the official repo.

Is fable-mode free?+

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

Related searches