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 agentmods add skills/casualsav/fable-bench/fable-methodnpx skills add casualsav/fable-bench --skill fable-methodgit clone --depth 1 https://github.com/casualsav/fable-benchWhat 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 | $0.00103 | $0.01468 |
| Opus 5 | $0.00051 | $0.00734 |
| Sonnet 5 | $0.00021 | $0.00294 |
| Haiku 4.5 | $0.00010 | $0.00147 |
Grade A, and why
fable-method 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 2d 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 — 128 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Fable method
The loop in one line: state the finish line → attack the riskiest assumption → build a thin end-to-end slice → verify through a channel you didn't build through → re-derive the plan from the goal → ship with verification status stated.
Everything below is that loop unpacked. Each rule is checkable — if you can't tell whether you followed it, you didn't.
1. Decomposing hard tasks
Restate the finish line before touching anything. Convert the ask into an observable end-state: "done when X passes / Y renders / Z returns this value." If you can't state it, the task is under-specified — resolving that IS the first sub-task, not a reason to guess.
Find the load-bearing uncertainty and test it first. Every hard task has one or two assumptions that, if wrong, invalidate everything downstream (the API supports this, the data has that shape, the library can do X). Name them out loud, then order the plan so the cheapest test of the riskiest assumption comes first. Plan order = risk order, never narrative order. An hour on step 3 is wasted if step 5 kills the approach.
Cut at seams you can state as contracts. A valid sub-task has an output you can specify in one sentence and check without redoing the work ("returns the list of endpoints with their auth requirements" — not "look into auth"). If you can't write a sub-task's acceptance check, it isn't decomposed yet; split along a different seam.
Separate the irreversible spine from the reversible flesh. Schema, API shape, public names, data migrations, anything sent externally — expensive to unwind; deliberate there. Internal code — cheap to revert; move fast there. Budget your care by cost-of-being-wrong, not by size-of-diff.
Prefer a thin end-to-end slice over layer-by-layer. Get one input flowing to one correct output through every layer first, then widen. Layer-by-layer defers all integration risk to the end, where it's most expensive; a thin slice converts unknown-unknowns into ordinary bugs on day one.
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.
- 2d ago First seen · 128 lines · 103 tokens per session scan A 8af807be7e61
fable-method is a skill published in the GitHub repository casualsav/fable-bench (5 stars, last pushed 5d ago), licensed MIT. It adds 103 tokens to every session and 1,468 once invoked, about $0.0005 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 2026-08-31.
Other skills, from other repositories
systematic-debugging
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.
brainstorming
You MUST use this before any creative work - creating features, building components, adding functionality, or modifying behavior. Explores user intent, requirements and design before implementation.
auto-perf-optimize
Run agent-driven VS Code performance or memory investigations. Use when asked to launch Code OSS, automate a VS Code scenario, run the Chat memory smoke runner, capture renderer heap snapshots, take workflow screenshots, compare run summaries, or drive a repeatable scenario before heap-snapshot analysis.
chat-perf
Run chat perf benchmarks and memory leak checks against the local dev build or any published VS Code version. Use when investigating chat rendering regressions, validating perf-sensitive changes to chat UI, or checking for memory leaks in the chat response pipeline.
chat-pet-sprite-creation
Use when creating or changing VS Code chat pet sprite art, sprite sheets, state animations, eye treatments, Stable/Insiders variants, or pet transitions under src/vs/workbench/contrib/chat/browser/widget/media/chatPet.
cpu-profile-analysis
Analyze V8/Chrome CPU profiles (.cpuprofile) and DevTools trace files (Trace-.json). Use when: profiling performance, investigating slow functions, comparing code paths, finding bottlenecks, analyzing timeToRequest, understanding call trees from sampling profiler data, analyzing layout/paint/rendering, investigating…