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 skills/cooltheworld/evolution-skills/steelmannpx skills add coolTheWorld/evolution-skills --skill steelmangit clone --depth 1 https://github.com/coolTheWorld/evolution-skillsWrote 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.
[](https://agentmods.dev/skills/cooltheworld/evolution-skills/steelman)<a href="https://agentmods.dev/skills/cooltheworld/evolution-skills/steelman"><img src="https://agentmods.dev/badge/skills/cooltheworld/evolution-skills/steelman.svg" alt="Measured on agentmods" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00022 | $0.00715 |
| Opus 5 | $0.00011 | $0.00358 |
| Sonnet 5 | $0.00004 | $0.00143 |
| Haiku 4.5 | $0.00002 | $0.00072 |
Grade A, and why
steelman 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 4d 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 — 62 lines — stays where its author put it; the contents beside it link to each section on GitHub.
A question arrives pre-framed, and the framing is usually the weakest part of it. Hold the answer: steelman both sides, isolate the crux, and rule only once the crux is settled.
Write in the language the user writes in.
1. Restate the real problem
Say back what the user is actually trying to solve, in a stronger and more complete form than they stated it — the version they would have written with more time.
Separate the stated ask from the underlying goal. Name the constraints they are treating as fixed, and what outcome would count as solved.
Done when the restatement names the goal, the fixed constraints, and the win condition, and is one the user would sign.
2. Steelman both sides
Build the strongest case for the user's current position, then the strongest case against it. Where they hold no position yet, steelman the two strongest live options instead.
Argue each side as its best advocate would — from its best evidence, granting it its most favourable reading. A case the opposing advocate would call unfair is a strawman wearing the label.
Look up whatever can be settled by fact — files, tools, the web — and argue from what you find rather than from what sounds plausible.
Done when each case rests on premises the other side concedes, and neither was easy to write.
3. Isolate the crux
The crux is where the two cases actually part company: the variable whose value, flipped, flips the verdict. Surface disagreement usually is not the crux — both sides often share the facts and diverge on one weighting, one probability, one constraint.
State where the cases genuinely diverge, then rank the crux variables by how far the conclusion moves with each.
Done when each crux reads as a conditional you can check: if it is X, the answer is A; if it is Y, the answer is B.
4. Ask the crux question
Ask the single question whose answer moves the verdict most. Ask exactly one, then stop and wait for the answer.
Carry your own recommended answer with the question, so the user is correcting a position rather than filling a blank.
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.
- 4d ago First seen · 62 lines · 22 tokens per session scan A 64df4f9ac684
steelman is a skill published in the GitHub repository coolTheWorld/evolution-skills (5 stars, last pushed 2d ago), licensed Apache-2.0. It adds 22 tokens to every session and 715 once invoked, about $0.0001 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 skills, from other repositories
skill-creator
Create or improve Zhin Agent skills (SKILL.md). Use when asked to add a skill, write SKILL.md, document a repeatable agent workflow, or refine skill frontmatter/keywords. Triggers: 创建技能, 写 SKILL, skill-creator, 加 skill.
webman
Expert skill for the webman framework (a long-lived, in-memory PHP framework based on workerman). Covers routing, controllers, middleware, database/Redis, custom processes, timers, coroutines (v2), plugin development, and guarding against memory leaks and cross-request state pollution under the resident process model.…
clawallex-skill
Pay for anything with USDC — virtual cards for any online checkout.
building-edgespark-apps
Build and modify EdgeSpark apps. Use when a project has edgespark.toml, the user mentions EdgeSpark, or work involves the edgespark CLI, server SDK types, storage/auth/database workflows, deployment, or @edgespark/web.
edgespark-frontend-design
Design and redesign EdgeSpark frontends with distinctive, production-grade visual direction instead of generic AI-looking UI. Use when building or polishing landing pages, marketing sites, dashboards, auth flows, portfolios, product surfaces, or reusable frontend sections in EdgeSpark, especially when the task…
iflytek-hyper-tts
Use when user asks to synthesize speech, convert text to audio, or read text aloud. 讯飞超拟人语音合成 - 支持文本转语音、语音合成（发音人/语速/语调/音量/输出格式）。大模型语音合成技能。语音合成, 文字转语音, 超拟人, TTS.