autoresearch SKILL.md skillA

autoresearch from dannykkh/skill-olympus is an agent skill: a SKILL.md folder an AI agent loads when a task calls for it.

Install it, pinned
mdr add dannykkh/skill-olympus/autoresearch@git:20261002.a341c48

Pins git:20261002.a341c48 in mdr.lock. Pin the fingerprint instead.

Audit grade
A, 17 of 17 checks passed
Kind
SKILL.md skill
Repository
dannykkh/skill-olympus
Path
skills/autoresearch/SKILL.md
Versions recorded
6
Latest version
git:20261002.a341c48, committed Oct 2, 2026
First version recorded
Mar 28, 2026
Size
27,212 bytes
Repository stars
7
License
MIT

Karpathy의 autoresearch 패턴을 스킬 프롬프트 최적화에 적용. 체크리스트(3~6개 예/아니오 기준) 기반 hill-climbing 루프로 SKILL.md를 자동 개선합니다. 한 번에 하나만 바꾸고, 채점하고, 유지하거나 되돌립니다. 95%+ × 3회 연속 달성 시 종료. 최종 합격은 라운드에 쓰지 않은 holdout 입력으로 독립 채점합니다. 기본 설치에서는 카탈로그의 source-only 원본을 직접 읽고, 전체 활성화 설치에서는 /autoresearch로 실행.

Collected from public GitHub. Every version is saved exactly as it was and never changes. Install it at an exact version with the mdr tool, and every file is checked before it reaches your agent. The main file is checked against the SHA-256 fingerprint recorded here, and the others against the fingerprints GitHub keeps for its source commit. The automatic checks below grade the latest version, and the same checks always give the same file the same grade.

What the file says

# Autoresearch

> **Karpathy의 autoresearch** — 한 번에 하나만 바꾸고, 채점하고, 유지하거나 되돌린다.
> 레시피를 고쳐서 앞으로 만드는 요리를 전부 좋게 만드는 것.

---

## 핵심 원리

```
점수를 매길 수 있으면, autoresearch할 수 있다.
— Ole (@unclejobs)
```

| 원본 (Karpathy ML) | 이 스킬 (프롬프트 최적화) |
|--------------------|-----------------------|
| `train.py` 수정 | `SKILL.md` 수정 |
| `program.md` 지침 | 이 SKILL.md = 루프 지침 |
| `val_bpb` 메트릭 | 체크리스트 점수 (pass/total × 100) |
| 학습 데이터 ≠ 검증 데이터 | 최적화 입력 ≠ holdout 입력 (holdout은 라운드에 쓰지 않고 최종 판정에만) |
| `results.tsv` | `autoresearch-log.md` 변경 로그 |
| `git commit/reset` | 유지/되돌림 |

---

## 사용법

기본 설치에서는 카탈로그에서 이 `SKILL.md`를 직접 읽고 아래 인자 의도를 적용합니다. 다음 표기는
명령 실행이 아니라 route intent입니다. `--include-source-only-skills`로 전체 활성화한 설치에서만
같은 앞에 `/`를 붙인 `/autoresearch` compatibility alias를 사용할 수 있습니다.

```text
# 기본 — 단일 스킬 최적화
autoresearch humanizer

# 옵션 지정
autoresearch humanizer --input sample.md --rounds 20

# 체크리스트 파일 직접 지정
autoresearch humanizer --checklist my-checklist.md

# 스캔 모드 — 전체 스킬 품질 진단
autoresearch --scan

# 스캔 후 바로 최적화
autoresearch --scan --auto
```

### 인자

| 인자 | 필수 | 기본값 | 설명 |
|------|------|--------|------|
| `<skill-name>` | — | — | 개선할 대상 스킬 이름 (scan 모드 시 생략) |
| `--scan` | — | — | 전체 스킬 품질 스캔 모드 |
…

Read the whole file at its exact version.

How to install

The exact contents, by fingerprint
mdr add dannykkh/skill-olympus/autoresearch@sha256:a54419afdaba9bf9

The line at the top of this page pins the label git:20261002.a341c48, which is what most people want. This one pins the sha256 fingerprint of the contents instead, so it survives a label being moved. Either way the exact version is written to mdr.lock, and mdr install downloads and checks those contents again, so it installs them exactly or stops with an error.

Badge

mdr badge

[![mdr](https://markdownregistry.com/badge/art_zb6cfjqguc4aguxj.svg)](https://markdownregistry.com/a/art_zb6cfjqguc4aguxj)

3 badge image fetches in 30 days. What the badge shows.

Versions

The registry has recorded 6 versions of this file, between Mar 28, 2026 and Oct 2, 2026, about one every 38 days.

versioncommittedcommitsizeaudit
git:20261002.a341c48 latest2026-10-02 a341c48 27,212 BA view · diff
git:20260901.8bc747e2026-09-01 8bc747e 21,799 BA view · diff
git:20260812.afc99502026-08-12 afc9950 21,767 BA view · diff
git:20260623.9b0dc322026-06-23 9b0dc32 20,696 BA view · diff
git:20260529.b363f962026-05-29 b363f96 19,590 BA view · diff
git:20260328.f05ffbb2026-03-28 f05ffbb 19,585 BA view

Audit of the latest version

A  17 of 17 checks passed. Fixed automatic checks with no AI involved, so the same checks give the same answer every run.
  • pass: Frontmatter block present
  • pass: Frontmatter declares a name
  • pass: Frontmatter declares a description
  • pass: Size between 200 bytes and 200 KB (27212 bytes)
  • pass: No zero-width or bidi control characters
  • pass: No instruction hidden inside an HTML comment
  • pass: No link to an exfiltration or paste host
  • pass: No credential-shaped string
  • pass: No instruction to send local credentials anywhere
  • pass: No text hidden with inline styles
  • pass: No prompt-injection phrasing
  • pass: No curl or wget piped into a shell
  • pass: No recursive delete of root, home or parent
  • pass: No instruction to read or print local credentials
  • pass: No base64 blob over 200 characters
  • pass: No link to a raw IP address
  • pass: No script tag

Source

GitHub

dannykkh/skill-olympus · 7 stars · license MIT · pushed 2026-10-22 · branch master

API

GET https://markdownregistry.com/api/v1/artifacts/art_zb6cfjqguc4aguxj
GET https://markdownregistry.com/api/v1/resolve?ref=dannykkh/skill-olympus/autoresearch
GET https://markdownregistry.com/api/v1/blob/a54419afdaba9bf9861deda964a89eb8c4130a8c7a1f46d573af15564f1d09ba

Your AI assistant does the legwork. You hear about the deals worth your word. Hand yours the standing instructions at modelranch.com and it joins the network that reads files like this one.

Guides: the SKILL.md format and its required fields · what to check before you install a skill · how to pin a skill to an exact version · where to find agent skills · keeping a team on the same skills · which agent file does what.

More from dannykkh/skill-olympus

AGENTS.md AGENTS.md file
dannykkh/skill-olympus · AGENTS.md
git:20261018.77243bb · audit A · 7 stars
DESIGN.md DESIGN.md file
dannykkh/skill-olympus · DESIGN.md
git:20260901.44bff6a · audit A · 7 stars
agent-md-refactor SKILL.md skill
dannykkh/skill-olympus · skills/agent-md-refactor/SKILL.md · Use when AGENTS.md, CLAUDE.md, or similar instruction files are bloated (300+ lines). Splits monolithic files into…
git:20260901.8bc747e · audit A · 7 stars
agent-team-codex SKILL.md skill
dannykkh/skill-olympus · skills/agent-team-codex/SKILL.md · Codex CLI의 내장 default, explorer, worker를 조합해 병렬 구현을 수행한다. zephermine 섹션 모드와 자유 모드를 지원하며 /agent-team, /poseidon…
git:20261022.6de5f4c · audit A · 7 stars
agent-team SKILL.md skill
dannykkh/skill-olympus · skills/agent-team/SKILL.md · zephermine 섹션 기반 네이티브 멀티에이전트 오케스트레이션. 의존성 분석, Wave 그룹핑, 파일 소유권 분리, 병렬 구현과 통합 검증을 수행한다. Claude, Codex, Antigravity…
git:20261022.6de5f4c · audit A · 7 stars
antigravity-mnemo SKILL.md skill
dannykkh/skill-olympus · skills/antigravity-mnemo/SKILL.md · Install, remove, or diagnose Mnemo long-term conversation memory for Google Antigravity CLI using its global Stop hook…
git:20261012.94ae631 · audit A · 7 stars
antigravity SKILL.md skill
dannykkh/skill-olympus · skills/antigravity/SKILL.md · Use Google Antigravity CLI (`agy`) for independent code or plan review, large-context analysis, and a second-model…
git:20260901.8bc747e · audit A · 7 stars
api-handoff SKILL.md skill
dannykkh/skill-olympus · skills/api-handoff/SKILL.md · Create backend/frontend API handoff docs for endpoints, frontend data needs, and cross-team data requirements.
git:20260427.85307e1 · audit A · 7 stars
api-tester SKILL.md skill
dannykkh/skill-olympus · skills/api-tester/SKILL.md · API 연동 테스트 스킬. 프론트-백엔드 통신 검증, 프록시 설정 확인, CORS, JWT 토큰, 에러 응답, 파일 업로드 테스트. "API 테스트해줘"로 실행.
git:20261002.d0b4bfd · audit A · 7 stars
argos SKILL.md skill
dannykkh/skill-olympus · skills/argos/SKILL.md · 준공검사 감리 스킬. 설계 산출물(spec, api-spec, qa-scenarios, flow-diagrams) 대비 구현을 정적·런타임·API·QA·도면·보안 관점으로 검증한다. 일반 diff 품질은 가용한…
git:20261022.6de5f4c · audit A · 7 stars
auto-continue-loop SKILL.md skill
dannykkh/skill-olympus · skills/auto-continue-loop/SKILL.md · 네이티브 /goal(Stop 게이트)을 녹여 쓰는 자율 작업 규율 스킬. 주 진입점은 /chronos. goal은 "끝까지 안 멈추기"(지속성)를 담당하고, 크로노스는 goal이 못 하는 것 — 자가 판단이 아닌…
git:20261020.a4ebf4e · audit A · 7 stars
biz-strategy SKILL.md skill
dannykkh/skill-olympus · skills/biz-strategy/SKILL.md · 사업성 검토 통합 스킬. 수요 검증 + 비즈니스 모델 캔버스, 수익/가격 전략, 시장 분석(TAM/SAM/SOM), GTM 전략, 북극성 지표, 코호트 분석을 한 번에 수행한다. /hermes, 헤르메스, 사업성…
git:20260812.afc9950 · audit A · 7 stars

Every file in dannykkh/skill-olympus

Other files named autoresearch

autoresearch SKILL.md skill
github/awesome-copilot · skills/autoresearch/SKILL.md · Autonomous iterative experimentation loop for any programming task. Guides the user through defining goals, measurable…
git:20260318.05f4d67 · audit A · 40,226 stars
autoresearch SKILL.md skill
yeachan-heo/oh-my-claudecode · skills/autoresearch/SKILL.md · Stateful single-mission improvement loop with strict evaluator contract, markdown decision logs, and max-runtime stop…
git:20260419.5c5835a · audit A · 40,148 stars
autoresearch SKILL.md skill
orchestra-research/ai-research-skills · 0-autoresearch-skill/SKILL.md · Orchestrates end-to-end autonomous AI research projects using a two-loop architecture. The inner loop runs rapid…
v1.0.0 · audit A · 13,718 stars
autoresearch SKILL.md skill
ericosiu/ai-marketing-skills · autoresearch/SKILL.md · Run Karpathy-style autoresearch optimization on any content. Generates 50+ variants, scores with a 5-expert simulated…
git:20260404.0076d34 · audit A · 3,676 stars
autoresearch SKILL.md skill
yeachan-heo/gajae-code · packages/coding-agent/src/defaults/gjc/skills/autoresearch/SKILL.md · Goal-directed research missions that interleave web research with data experimentation and end on a structured…
git:20260823.9983f32 · audit A · 3,016 stars
autoresearch SKILL.md skill
mxyhi/ok-skills · autoresearch/SKILL.md · Autonomous iteration loop: modify, verify, keep/discard against any metric
v2.2.2 · audit A · 504 stars
autoresearch SKILL.md skill
naveedharri/benai-skills · plugins/autoresearch/skills/autoresearch/SKILL.md · Autonomous goal-directed iteration loop that continuously improves prompts, templates, configs, or code. Two evaluation…
git:20260916.a3b1e15 · audit A · 71 stars
autoresearch SKILL.md skill
naveedharri/benai-skills · plugins/all-skills/skills/autoresearch/SKILL.md · Autonomous goal-directed iteration loop that continuously improves prompts, templates, configs, or code. Two evaluation…
git:20260916.a3b1e15 · audit A · 71 stars

Browse by kind, by grade A, or by owner.