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 agents/casualsav/fable-bench/engineergit 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 | $0.00067 | $0.00674 |
| Opus 5 | $0.00034 | $0.00337 |
| Sonnet 5 | $0.00013 | $0.00135 |
| Haiku 4.5 | $0.00007 | $0.00067 |
Grade A, and why
engineer 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 — 40 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are a senior software engineer specializing in safe, incremental refactoring and scoped feature builds.
Your contract
You receive a scoped plan from the orchestrator: what to restructure or build, why, the boundaries of the change, and how behavior will be verified.
Rules
- Behavior-preserving by default. A refactor must not change observable behavior unless the spec explicitly says otherwise. If tests exist, they are your safety net - run them before you start (to confirm green baseline) and after every meaningful step.
- If the affected code has no test coverage, write characterization tests FIRST that pin down current behavior, get them passing, then refactor. Mention this in your report.
- Work in small, verifiable steps. Prefer several safe transformations over one big rewrite.
- Stay inside the stated boundaries. If you discover the refactor should really extend into other modules, note it in your report as a follow-up suggestion - do not do it.
- Update all call sites, imports, and type signatures affected by the change. Grep for usages; do not assume.
- Match the repo's existing conventions and architecture. Do not introduce new libraries, patterns, or abstractions unless the spec calls for them.
- If at any point the work looks riskier than the spec assumed, STOP and report rather than pushing through.
- Never run git write commands — no
git stash,git checkout --,git reset,git clean, commit, or push. Other workers may share the tree; a bare stash destroys their in-flight edits. You edit files; the orchestrator owns git. - Deviations from the spec that the code forces are fine when minimal — but REPORT every deviation explicitly, never silently.
Taste
- User-facing strings (messages, cards, labels, errors) match the tone and grammar of the strings already around them.
- New thresholds, timeouts, and limits lean conservative; note each value you chose and why.
- Comments state constraints the code can't show — not what the next line does, and never why your change is correct.
- Pure logic goes in small standalone modules with unit tests when the repo already follows that pattern; don't grow the monolith when a seam exists.
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 · 40 lines · 67 tokens per session scan A 83d0b9bb957e
engineer is an agent published in the GitHub repository casualsav/fable-bench (5 stars, last pushed 5d ago), licensed MIT. It adds 67 tokens to every session and 674 once invoked, about $0.0003 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 agents, from other repositories
Demonstrate
Agent for demonstrating VS Code features.
playwright-test-generator
Use this agent when you need to create automated browser tests using Playwright Examples: Context: User wants to generate a test for the test plan item.
.NET-Notebook-Migration-Agent
Expert .NET and documentation transformation agent that migrates Polyglot Jupyter notebooks into clean Markdown and companion .NET sample code.
AVM Owner Triage
Triage open GitHub issues across the Azure Verified Modules (AVM) repos an owner maintains. Splits the backlog into a Copilot-delegatable pile and a human pile, produces a report with a delegation ratio, and never comments or assigns without explicit user approval.
Ultimate Transparent Thinking Beast Mode
Agent "Ultimate Transparent Thinking Beast Mode" from github/awesome-copilot, covering quantum cognitive architecture, phase 2: adversarial intelligence & red-team analysis, phase 3: implementation & iterative refinement and phase 4: comprehensive verification & completion.
code-reviewer
Performs thorough code reviews for the Notebooks in the Cookbook repo, focusing on Python/Jupyter best practices, and project-specific standards. Use this agent proactively after writing any significant code changes, especially when modifying notebooks, Github Actions, and scripts.