Skip to main content
Back to the directory
alchaincyf/darwin-skillSoftware EngineeringFrontend and Design

darwin-skill

SkillJury is still gathering richer source copy for this listing. Community reviews, install context, and source proof are shown below.

SkillJury keeps community verdicts, source metadata, and external repository signals in separate lanes so ranking data never pretends to be a review.

SkillJury verdict
Pending

No approved reviews yet

Would recommend
Pending

Waiting on enough review volume

Install signal
3

Weekly or total install activity from catalog data

Sign in to review
0 review requests
Install command
npx skills add https://github.com/alchaincyf/darwin-skill --skill darwin-skill
SkillJury does not have enough approved reviews to publish a community verdict yet. Source metadata and repository proof are still available above.
SkillJury Signal Summary

As of Apr 30, 2026, darwin-skill has 3 weekly installs, 0 community reviews on SkillJury. Community votes currently stand at 0 upvotes and 0 downvotes. Source: alchaincyf/darwin-skill. Canonical URL: https://skills.sh/alchaincyf/darwin-skill/darwin-skill.

Security audits
Gen Agent Trust HubPASS
SocketPASS
SnykWARN
About this skill
借鉴 Karpathy autoresearch 的自主实验循环,对 skills 进行持续优化。 核心理念: 评估 → 改进 → 实测验证 → 人类确认 → 保留或回滚 → 生成成果卡片 GitHub: autoresearch 的精髓: 与纯结构审查的区别:不只看 SKILL.md 写得规不规范,更看改完后 实际跑出来的效果是否更好 。 这是与纯结构评分最大的区别。评分方式: 如果无法跑子agent(时间/资源限制),可以退化为「干跑验证」:读完skill后模拟一个典型prompt的执行思路,判断流程是否合理。但要在results.tsv中标注 dry_run 。 在评估之前,为每个skill设计测试prompt。这步很关键——没有测试prompt,「实测表现」维度就打不了分。 展示所有测试prompt给用户, 确认后再进入评估 。测试prompt的质量决定了优化方向是否正确。 如果子agent不可用 (超时、环境限制),维度8用干跑验证打分,标注 dry_run 。不要因为跑不了测试就跳过这个维度——哪怕是模拟推演也比完全不看效果好。 基线评估完成后,展示评分卡: 暂停等用户确认,再进入优化循环。 用户确认后,按基线分数从低到高排序,先优化最弱的。 当 hill-climbing 连续2个skill都在 round 1 就 break(涨不动)时,提议一次「探索性重写」: 这解决了 hill-climbing 的局部最优问题——有时候需要「先拆后建」才能突破瓶颈。 必须征得用户同意后才执行。 新增 eval_mode 列: full_test (跑了子agent测试)或 dry_run (模拟推演)。 文件位置: .claude/skills/darwin-skill/results.tsv 按优先级排序,每轮只做最高优先级的一个: 流程假设环境理想,但实操常遇异常。以下预定义 fallback,保证优化过程不会「一跑就卡住」。 原则 :异常先告知用户,再按规则处理;绝不静默跳过或静默失败。 "You write the goals and constraints in program.md; let an agent generate and test code deltas indefinitely; keep only what measurably improves the objective." — Karpathy, autoresearch 本skill的对应关系: 区别:增加了人在回路(autoresearch是全自主的,skill优化需要人的判断力),以及双重评估机制(结构+效果),因为skill的「好坏」比loss数值更微妙。 每个skill优化完成后(或全量汇总后),自动生成视觉成果卡片,截图保存为PNG。 模板位置: templates/result-card.html 3种风格,每次随机选择一种: - 单一可编辑资产 — 每次只改一个 SKILL.md - 双重评估 — 结构评分(静态分析)+ 效果验证(跑测试看输出) - 棘轮机制 — 只保留改进,自动回滚退步 - 独立评分 — 评分用子agent,避免「自己改自己评」的偏差 - 人在回路 — 每个skill优化完后暂停,用户确认再继续 - 维度1-7:每个维度打...

Source description provided by the upstream listing. Community review signal and install context stay separate from this narrative layer.

Community reviews

Latest reviews

No community reviews yet. Be the first to review.

Browse this skill in context
FAQ
What does darwin-skill do?

darwin-skill is listed in SkillJury, but the source summary is still sparse.

Is darwin-skill good?

darwin-skill does not have approved reviews yet, so SkillJury cannot publish a community verdict.

Which AI agents support darwin-skill?

darwin-skill currently lists compatibility with Skills CLI.

Is darwin-skill safe to install?

darwin-skill has been scanned by security audit providers tracked on SkillJury. Check the security audits section on this page for detailed results from Socket.dev and Snyk.

What are alternatives to darwin-skill?

Skills in the same category include grimoire-morpho-blue, conversation-memory, second-brain-ingest, zai-tts.

How do I install darwin-skill?

Run the following command to install darwin-skill: npx skills add https://github.com/alchaincyf/darwin-skill --skill darwin-skill

Related skills

Alternatives in Software Engineering