fairy_tale: Skill for Claude Code

.agents/skills/fairy-tale-benchmark-feedback/SKILL.md

fairy-tale-benchmark-feedback is a skill for Claude Code, Codex from bonginkan/fairy_tale. It costs 0 tokens per session (2,302 once invoked), scanned A, original, Apache-2.0.

A feedback process for improving coding-agent behaviour after benchmark failures or successful test runs, using visible evidence and reusable rules.

In plain words
What is it for?
Use it to record feedback for SWE-Bench Pro, HLE-style, or defensive ExploitBench runs and retry selected samples with the revised guidance.
Why use it?
It helps identify the first actionable mistake, correct misleading guidance, and add narrow coverage without relying on hidden answers or scorer details.

Skill for Claude CodeCodex

Written for no agent in particular: nothing here depends on one. Also seen: installed under .agents/ (shared by several agents).

This is bonginkan/fairy_tale's own configuration. It tells Claude Code and Codex how to work on fairy_tale itself, so it is not a mod to install elsewhere. Copy it as a starting point and replace the rules that are about this project. Everything fairy_tale configures →

Reuse

Borrowing it

Nothing to install: this file belongs to bonginkan/fairy_tale. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file
curl -O https://raw.githubusercontent.com/bonginkan/fairy_tale/main/.agents/skills/fairy-tale-benchmark-feedback/SKILL.md
Clone the repo
git clone --depth 1 https://github.com/bonginkan/fairy_tale

Made for: Claude Code, Codex.

Wrote 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.

agentmods badge for fairy-tale-benchmark-feedback

README.md
[![agentmods](https://agentmods.dev/badge/skills/bonginkan/fairy_tale/fairy-tale-benchmark-feedback/github.svg)](https://agentmods.dev/skills/bonginkan/fairy_tale/fairy-tale-benchmark-feedback)
Your own site
<a href="https://agentmods.dev/skills/bonginkan/fairy_tale/fairy-tale-benchmark-feedback"><img src="https://agentmods.dev/badge/skills/bonginkan/fairy_tale/fairy-tale-benchmark-feedback/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for fairy-tale-benchmark-feedback

Your own site · 80×15
<a href="https://agentmods.dev/skills/bonginkan/fairy_tale/fairy-tale-benchmark-feedback"><img src="https://agentmods.dev/badge/skills/bonginkan/fairy_tale/fairy-tale-benchmark-feedback.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 0 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 2,302 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe. Third-party audits
  • NVIDIA SkillSpector pass 7 Sept 2026
How audits are shown
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5.1 $0.00000 $0.02302
Opus 5 $0.00000 $0.01151
Sonnet 5 $0.00000 $0.00460
Haiku 4.5 $0.00000 $0.00230

Measured 11d ago against content hash ea77049234d7, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-11, from the pricing page.

Security

Grade A, and why

fairy-tale-benchmark-feedback 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 11d 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.

.agents/skills/fairy-tale-benchmark-feedback/SKILL.md · 205 lines

How it starts

The opening of the file, as written. The whole thing — 205 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Fairy Tale Benchmark Feedback

Use this skill after a measured benchmark miss, work-product failure, or successful benchmark slice whose practice should be made reproducible in agentic coding, HLE-style closed-ended reasoning, or defensive ExploitBench sandbox runs.

Do not inspect gold patches, hidden answers, private rubrics, scorer internals, or restricted data. Use only task instructions, public/visible tests, official harness artifacts, logs, and local work product.

Feedback Loop

  1. Preserve the run conditions: model, effort, prompt, tools, scorer, sample IDs, budget, concurrency, and artifacts.
  2. Convert misses and observed success practices into a feedback ledger:
    • SWE-Bench Pro: scripts/benchmark_feedback_ledger.py swe-bench-pro
    • HLE-style tasks: scripts/benchmark_feedback_ledger.py hle
    • ExploitBench: scripts/benchmark_feedback_ledger.py exploitbench
  3. Before writing a candidate rule, localize the first actionable fault step in the failed trajectory or work product. Link it to an existing Fairy Tale rule when that rule misled the run; otherwise mark it as missing coverage.
  4. Revise an existing rule when responsibility is clear. Generate a new narrow rule only when no existing rule can be safely revised. Make no skill update when the trace does not support the attribution.
  5. Run pruning before promotion: scripts/feedback_pruner.py --ledger <ledger.json> --output <prune.json>.
  6. Keep only narrow candidate or observed-success rules with evidence. Do not promote a rule because it sounds plausible.
  7. If the same failure signature repeats, a run produces no meaningful artifact, or the validation ledger is missing, run bounded Fairy Fusion before retry: isolated reviewers, one synthesis pass, append-only review artifacts, and only a compact closure hint returned to the main agent. Continue retrying until the local clear condition is met or the user/operator stops the run.
  8. Retry a held-out or failed slice under the same scorer. Record before/after pass rate, confidence interval when applicable, cost, and regressions.
  9. Promote only rules that improve the retry without task-ID hardcoding or cross-domain regression.

Read the full file on GitHub · 205 lines

Changes

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.

  1. 11d ago First seen · 205 lines · 0 tokens per session scan A ea77049234d7

Subscribe to this mod's changes

fairy-tale-benchmark-feedback is a skill published in the GitHub repository bonginkan/fairy_tale (18 stars, last pushed 3d ago), licensed Apache-2.0. It costs nothing until one of its globs matches a file; then it loads 2,302 tokens. 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.

Related

Other skills, from other repositories

research-engineer

An uncompromising Academic Research Engineer. Operates with absolute scientific rigor, objective criticism, and zero flair. Focuses on theoretical correctness, formal verification, and optimal implementation across any required technology.

davila7/claude-code-templates · 43 tokens

tika-eval-compare

Compare extracts from two Tika builds over a corpus to detect regressions in content, encoding, exceptions, and embedded-document handling. Use for "compare before/after extracts", "eval this change against the corpus".

apache/tika · 50 tokens

neuron-evaluation-engineer

Create and run AI evaluations with datasets, assertions, and output drivers in Neuron AI. Use this skill whenever the user mentions evaluation, testing AI systems, creating evaluators, dataset-driven testing, assertion-based validation, or wants to measure AI system performance. Also trigger for tasks involving…

neuron-core/neuron-ai · 77 tokens

jetson-validate-image

Use after jetson-flash-image to run static BSP checks, on-target smoke/regression tests on a flashed DUT, or both. Not for build or flash steps. Triggers: validate bsp, on-target validation.

NVIDIA/skills · 50 tokens

atmos-validation

Validate Atmos projects, components, arbitrary JSON Schema inputs, EditorConfig, and GitHub Actions; use affected-file selection and native CI annotations.

cloudposse/atmos · 31 tokens

skill-benchmark

Benchmark AI skill effectiveness by measuring implementation quality against legacy constraints.

HoangNguyen0403/agent-skills-standard · 16 tokens