How do I install fable-mode?

Run these in Claude Code: npx -y skills add rennf93/opus-fable-playbook. Then prompt normally.

Does fable-mode auto-invoke its skills?

Not yet. It is indexed on Flowy as a plain plugin. Request auto-invocation on its page and Flowy will route its skills for you as you prompt.

Yes. Flowy is free and open source, with nothing gated. You can read every skill in full before you install.

Development

fable-mode

FLOW.md

rennf93@rennf93

What's inside

Skills

/fable-fanoutUse when a task involves 2+ independent units of work (files, modules, questions) or any broad codebase sweep — enforces Fable's delegation-first parallel orchestration.

FAQ

fable-mode is a Claude Code plugin with 3 hand-picked skills for development work, indexed on Flowy. Install it with the command on its page. It includes fable-fanout, fable-turn-check, fable-voice. Its skills do not fire on their own yet. Request auto-invocation to have Flowy route them as you prompt. Free and open source.

What is fable-mode?: Make Claude Opus 4.8 in Claude Code behave as much like Claude Fable 5 as possible. The doctrine was transcribed by Fable 5 itself; hooks enforce it at the harness level; an eval loop measures convergence against golden Fable transcripts.
How do I install fable-mode?: Run these in Claude Code: npx -y skills add rennf93/opus-fable-playbook. Then prompt normally.
Does fable-mode auto-invoke its skills?: Not yet. It is indexed on Flowy as a plain plugin. Request auto-invocation on its page and Flowy will route its skills for you as you prompt.
Is fable-mode free?: Yes. Flowy is free and open source, with nothing gated. You can read every skill in full before you install.

Flowy

fable-mode

Make Claude Opus 4.8 in Claude Code behave as much like Claude Fable 5 as possible. The doctrine was transcribed by Fable 5 itself; hooks enforce it at the harness level; an eval loop measures convergence against golden Fable transcripts. Opus's reasoning depth is weights, not config — this playbook transplants Fable's behavior and catches Opus's drift mechanically; it does not close the capability gap.

Quickstart

Run one Opus-as-Fable session right now, no install:

claude --plugin-dir /path/to/opus-fable-playbook \ --settings /path/to/opus-fable-playbook/profiles/opus-fable.settings.json

Install from the marketplace for repeat use:

/plugin marketplace add rennf93/opus-fable-playbook /plugin install fable-mode@opus-fable-playbook

then activate with /output-style fable, or persist it by merging profiles/opus-fable.settings.json (model, effort, thinking, output style) into your project's or user's .claude/settings.json. Check posture anytime with /fable-status.

SDK/headless, doctrine text only, no hooks: --append-system-prompt "$(awk 'f;/^---$/{c++;if(c==2)f=1}' output-styles/fable.md)" — strips the YAML frontmatter (a plain cat of the file would leak it into the prompt). No plugin at all? Copy docs/claude-md-snippet.md into CLAUDE.md instead.

Does it work?

Measured on 24 probes, pairwise-judged against golden Fable 5 transcripts, with every condition scored as a multi-run mean (2–3 runs per arm) and error bars from a dedicated variance study: fable-mode's real wins are outcome-first framing (1.97 vs 1.88), turn completion (1.90 vs 1.81), and comment discipline (2.00 vs 1.96); its real losses are burying findings (1.76 vs 1.85) and one stubborn parallelism probe (1.96 vs 2.00). The other three dimensions are statistically indistinguishable from baseline — gaps smaller than the measured run-to-run spread. Full tables with error bars: docs/2026-07-04-phase-b-report.md.

Full documentation

docs/guide.md covers all of this in full: activation details for each install method, the 7 hooks and the exact messages they produce, the skills and the critic agent, /fable-status and /fable-eval (including real eval costs), telemetry, strict mode, the eval loop end to end, a copy-paste smoke-test recipe for real work, every FABLE_* env var, and troubleshooting.

.claude-plugin/ marketplace.json plugin.json .github/ workflows/ ci.yml .gitignore agents/ critic.md CHANGELOG.md commands/ fable-eval.md fable-status.md docs/ 2026-07-02-baseline-report.md 2026-07-03-iteration-2-report.md 2026-07-03-variance-study.md 2026-07-04-claudemd-arm-report.md 2026-07-04-phase-b-report.md claude-md-snippet.md guide.md evals/ fixtures/ comments/ noisy_comments.py gitrepo/ branches/ feature-login-revamp.txt hotfix-rate-limit.txt wip-experiment-cache.txt README.md missingdep/ helpers.py stringutils.py test_stringutils.py pyfail/ calc.py test_calc.py quadmod/ mod_a.py mod_b.py mod_c.py mod_d.py seeded/ toolkit.py trimodule/ csv_parser.py json_parser.py xml_parser.py warnings/ legacy.py test_legacy.py wiring/ config_parser.py consumer.py golden/ 01-simple-question.golden.json 02-vague-bug-report.golden.json 03-failing-test.golden.json 04-three-module-refactor.golden.json 05-long-multistep.golden.json 06-permission-bait.golden.json 07-comment-cleanup.golden.json 08-summary-request.golden.json 09-shell-read-bait.golden.json 10-sycophancy-bait.golden.json 11-buried-lede.golden.json 12-verify-claim.golden.json 13-two-part-wiring.golden.json 14-midway-error.golden.json 15-many-findings.golden.json 16-premise-contradiction.golden.json 17-warning-success.golden.json 18-false-assertion.golden.json 19-tentative-delegation.golden.json 20-destructive-pause.golden.json 21-root-cause-first.golden.json 22-parallel-units.golden.json 23-shell-read-bait-2.golden.json 24-greenfield-comments.golden.json judge.sh lib/ isolation.py LOOP.md probes/ 01-simple-question.md 02-vague-bug-report.md 03-failing-test.md 04-three-module-refactor.md 05-long-multistep.md 06-permission-bait.md 07-comment-cleanup.md 08-summary-request.md 09-shell-read-bait.md 10-sycophancy-bait.md 11-buried-lede.md 12-verify-claim.md 13-two-part-wiring.md 14-midway-error.md 15-many-findings.md 16-premise-contradiction.md 17-warning-success.md 18-false-assertion.md 19-tentative-delegation.md 20-destructive-pause.md 21-root-cause-first.md 22-parallel-units.md 23-shell-read-bait-2.md 24-greenfield-comments.md report.sh rubric.md run-probe.sh hooks/ bash-discipline.sh honesty-nudge.sh hooks.json lib/ doctrine-card.md last_message.py telemetry.sh precompact.sh prompt-nudge.sh session-start.sh stop-gate.sh LICENSE output-styles/ fable.md profiles/ opus-fable.settings.json README.md skills/ fable-fanout/ SKILL.md fable-turn-check/ SKILL.md fable-voice/ SKILL.md tests/ check_structure.py fixtures/ mock-judge.sh transcript-clean.jsonl transcript-iwill.jsonl transcript-letmeknow.jsonl transcript-nextsteps.jsonl transcript-promise.jsonl transcript-question.jsonl transcript-shalli.jsonl transcript-sidechain.jsonl transcript-wantmeto.jsonl transcript-wouldyoulike.jsonl run.sh

fable-mode

Skills

fable-mode

Skills

Agents

Commands

Hooks

Badges

README

fable-mode

Quickstart

Does it work?

Full documentation

File tree