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.
npx skills add Nam-Cheol/codex-fable-mode --skill fable-modegit clone --depth 1 https://github.com/Nam-Cheol/codex-fable-modeWrote 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.
[](https://agentmods.dev/skills/nam-cheol/codex-fable-mode/fable-mode)<a href="https://agentmods.dev/skills/nam-cheol/codex-fable-mode/fable-mode"><img src="https://agentmods.dev/badge/skills/nam-cheol/codex-fable-mode/fable-mode.svg" alt="Measured on agentmods" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00047 | $0.02614 |
| Opus 5 | $0.00023 | $0.01307 |
| Sonnet 5 | $0.00009 | $0.00523 |
| Haiku 4.5 | $0.00005 | $0.00261 |
Grade A, and why
fable-mode 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 7d 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.
How it starts
The opening of the file, as written. The whole thing — 247 lines — stays where its author put it; the contents beside it link to each section on GitHub.
fable-mode
fable-mode is a single operating mode for intent-aware work.
It is not a harness. It is not a rigid spec generator. It is not a design-only mode. It is not "think as much as possible."
The purpose of fable-mode is to choose the smallest sufficient reasoning depth and procedure that protects user intent, explicit constraints, grounding, capability fit, output form, correctness, and output quality.
fable-mode is domain-neutral. It must not push work toward one subject, audience, visual style, or artifact type. HTML, CSS, JavaScript, canvas, SVG, images, documents, tests, terminal commands, and prose are implementation tools. The desired output form comes from the user's intent.
fable-mode is not a Fable 5 clone, model bridge, provider bridge, runtime harness, automation framework, or goal engine. It does not unlock, emulate, proxy, or route to another provider. It is documentation-only guidance for Codex behavior.
Core loop
Before planning, using tools, or acting:
- Frame the user's intent.
- Lock the primary output type.
- Choose a depth level.
- Set a procedure budget.
- Extract explicit constraints.
- Decide whether to ask or act.
- Classify the requested output form.
- Identify grounding requirements.
- Check capability fit.
- Set a tool budget.
- Set a delegation budget.
- Decide whether to expose a compact route line.
- Infer audience intent when the user gives one.
- Use audit lanes only when required.
- Implement or answer.
- Run a compact pre-final critique.
- Report only what matters.
This is a routing loop, not a requirement to read every reference file.
The output lock must be one of: answer, edit, implementation, review, audit, design artifact, research, clarification, or validation. It determines reasoning depth, which reference docs to read, whether to ask or act, whether tools are allowed, how much verification is enough, and final response length.
Grounding and capability checks are stop gates when they protect a hard requirement. If a required source, asset, measurement, reproduction, runtime, browser, current lookup, or other capability is unavailable, do not enter implementation by substituting a plausible approximation. Stop, inspect with an available tool, ask for the missing material, or change the output to a clearly limited answer or plan.
What ships with it
44 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- agents/openai.yaml 43 B
- evals/delegation-budget.md 519 B
- evals/grounding-stop-gate.md 451 B
- evals/output-drift.md 499 B
- evals/over-scope.md 407 B
- evals/renderable-verification.md 507 B
- evals/route-disclosure.md 451 B
- evals/tool-bloat.md 389 B
- examples/a1-debugging.md 898 B
- examples/a2-review.md 798 B
- examples/l0-route-omission.md 818 B
- examples/l2-implementation-route.md 722 B
- examples/l3-design-route.md 801 B
- references/ask-or-act.md 1.3 KB
- references/audience-intent.md 1.1 KB
- references/audit-lanes.md 2.2 KB
- references/behavior-model.md 1.8 KB
- references/behavioral-evaluation.md 1.0 KB
- references/capability-fit.md 2.4 KB
- references/coding-workflow.md 1.9 KB
- references/constraint-integrity.md 1.3 KB
- references/delegation-budget.md 2.1 KB
- references/depth-gate.md 1.9 KB
- references/design-anti-patterns.md 3.0 KB
- references/design-exploration.md 2.7 KB
- references/design-review-rubric.md 2.7 KB
- references/design-thinking.md 1.9 KB
- references/final-response.md 822 B
- references/grounding-integrity.md 2.2 KB
- references/intent-framing.md 1.2 KB
- references/investigation-protocol.md 1.2 KB
- references/output-archetype.md 2.3 KB
- references/output-form-integrity.md 1.4 KB
- references/output-lock.md 4.0 KB
- references/pre-final-critique.md 852 B
- references/procedure-budget.md 3.2 KB
- references/run-card.md 1.0 KB
- references/safety-boundaries.md 1.8 KB
- references/small-fix-protocol.md 1.4 KB
- references/substrate-selection.md 4.5 KB
- references/tool-budget.md 3.0 KB
- references/tool-policy.md 2.2 KB
- references/verification-gate.md 1.2 KB
- references/visible-route-contract.md 2.0 KB
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.
- 7d ago First seen · 247 lines · 47 tokens per session scan A a0e6cdf0a070
fable-mode is a skill published in the GitHub repository Nam-Cheol/codex-fable-mode (27 stars, last pushed 2mo ago), licensed MIT. It adds 47 tokens to every session and 2,614 once invoked, about $0.0002 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 2026-08-30.
Other skills, from other repositories
cert-study
A study session for SQLD, a SQL database certification exam, run inside Codex with SQLite records and error notes for Obsidian.
watch
File sentinel that monitors the working directory for changes and marker comments, then auto-triggers appropriate skills. Poll-based via git diff against the last scan commit. Writes intake items for batch processing and routes marker actions through /do. Use for automatic reactions to file changes; do NOT use for…
live-preview
Mid-build visual verification loop. Takes screenshots of components during construction, not just after. Catches visual regressions and invisible features before they compound. Requires Playwright or similar screenshot tool.
review
5-pass structured code review — correctness, security, performance, readability, consistency.
marshal
Meta-orchestrator that takes any direction — broad, specific, or vague — and autonomously chains skills and context into actionable work. Gathers context from codebase, docs, and memory. Only asks the user when it genuinely cannot proceed. Single-session orchestrator.
wiki
Markdown-first knowledge base where the LLM acts as librarian. Ingests raw sources, compiles and interlinks topic files, self-maintains an index. No vector DB or embeddings required -- uses LLM-native navigation over structured markdown up to 400K words.