fable-method
The Fable Workflow: how Claude Fable 5 worked, distilled into skills any model can run, with the eval that keeps it honest. Think / act / prove.
Runssource: GitHubPythonMITcommit 88b5cf36b10e
Python, MIT licensed. The project labels itself: agent skills, ai agents, claude, claude code, claude md, coding agent, evaluation and fable.
fable-method runs. An Argusic agent installed it in 1 minute and hit no errors on a clean machine with no GPU, and the whole session was recorded.
Repo is a documentation/skill distribution: the install script copies skill files to $HOME/.claude/skills, the repo checks pass (50/50), and all 14 eval trap scenarios work as documented, no dependencies, no build step, no runtime beyond Python 3 and Node.js which are both available.
What the agent ended up with on a clean machine, in its own words. How this is measured
Measured by Argusic on a fresh machine every time. Every number links to its evidence. Argusic Score 100 of 100 (the mean of this project's run scores).
At a glance
Subject data from GitHub, linked at the top of this page, refreshed . Test data by Argusic (CC BY 4.0); every number links to a run page with the full log, the recording, and their sha256 hashes.
What was measured
From the run that decided this verdict. Every line is a recorded value, not an estimate.
- Install time
- 1 minutes
- Cold machine to finish
- 2 minutes
- Errors hit and fixed
- 0 hit, 0 fixed with no human help
- How the result was proved
- python3 .github/checks.py exited 0 with 50 checks passed; install.sh copied 3 skills to ~/.claude/skills/; all 14 eval scenarios exercised and verified against their ground truth (S2 surprise trap: test fails as designed, S3 UTC bucketing: report runs and outputs 3 daily counts, S10 recall trap: fetchlite client returns 4 of 7 users active, S1 assessment trap: calculateTotal produces 19.60 instead of 19.44 confirming the tax-before-discount bug, S9 unauthorized action: check_config.py correctly reports db_path and timeout_seconds errors, S4 messy export: CSV analysis yields Gadget Max 640 > Widget Pro 550 > Doohickey 175)
- Model tokens used
- 42,190
- Exact commit tested
- 88b5cf36b10e
- Container image
- sha256:cdd920bce783
What was not measured
The other half, so you can see exactly how far this evidence goes. A gap is printed as a gap and never filled in.
- Peak GPU memory
- this ran on a CPU machine, so there was no GPU to sample
- Behaviour over a long run
- each attempt uses the smallest documented settings and a fixed time budget, so this measures whether it runs, not how it holds up under load
The end of the recording
The last lines of the terminal session that decided this verdict, exactly as they were recorded.
2026-11-11T16:19:12.929943Z ERROR codex_core::tools::router: error=exec_command failed: CreateProce… 2026-11-11T16:19:14.799442Z ERROR codex_core::tools::router: error=exec_command failed: CreateProce… exec /bin/sh -lc 'cd /work/repo && rm -r eval/scenarios/s10-recall-trap/lib/__pycache__/ eval/scenarios/… succeeded in 0ms: exec /bin/sh -lc 'cd /work/repo && git status --short 2>/dev/null' in /work/repo succeeded in 0ms: Clean. Here's the summary: tokens used 42,190 Clean. Here's the summary:
Replay the whole session, every command from a clean machine to this point.
How it was tested
One agent, one container, no human help. It installs the project the way its own documentation says to, runs it, and fixes what breaks. Everything below is recorded as it happened: the terminal session, the log and the exact commit. The full method.
Strengths and limits
Measured facts, not opinions. How this is written.
What went well
- Reached a running state on a clean machine, with the session recorded.
- Installed in 1 minute, faster than the median of the 122 comparable projects Argusic has measured.
- Nothing broke on the way: zero errors between clone and running.
- Ran without a GPU, so it does not need one to start.
- Licensed MIT, as reported by its host.
- Measured 2 times, so the result is not a one-off.
Of the 123 agent-skills projects Argusic has installed and timed, fable-method was the 31st fastest to reach a running state, and 94 of 123 reached one at all.
What it is a reasonable choice for
- Trying it on a laptop or a small server: it reached a running state without a GPU.
- A quick evaluation: it was running 1 minutes after the clone.
- Following its own README: the documented path worked with nothing to fix.
Also tested, in the same area
Every one of these was installed and run by Argusic on a clean machine. Nothing appears here that was not tested.
Run history
Topics (from GitHub)
agent-skillsai-agentsclaudeclaude-codeclaude-mdcoding-agentevaluationfablellm
Embed the badge
Markdown for the project README. It links back here; terms on the terms page.
[](https://argusic.com/subject/fable-method)Questions
- Does fable-method run?
- Yes. fable-method runs. Argusic installed and launched it on a clean machine in 1 minutes, hitting 0 errors on the way, and recorded the session.
- How did Argusic test fable-method?
- On a fresh, disposable machine, with every command recorded and the repository pinned at commit 88b5cf36b10e. 2 attempts are recorded, and the full method is on the methodology page.
- What did it cost to test fable-method?
- The run that produced this verdict cost $0.03: real compute and model cost, not a list price per million tokens. It is recorded on that run's page.
- How long does fable-method take to install?
- 1 minute in the run that decided this verdict, from a clean machine to a running state. That is one measured attempt, not an average.
- Does fable-method need a GPU?
- Not to start. Argusic reached a running state on a machine with no GPU.
- What goes wrong when installing fable-method?
- Nothing did, in the recorded run: zero errors between clone and running.
- How does fable-method compare with the alternatives?
- Of the 123 agent-skills projects Argusic has installed and timed, fable-method was the 31st fastest to reach a running state, and 94 of 123 reached one at all.
- Where is the evidence for fable-method?
- All 2 recorded runs are on this page, each linking to its full log and terminal recording, stored with a sha256 fingerprint so it cannot be quietly altered.