criptogus/agent-evolve-network

최근 기록된 소스 활동: 2026년 10월 25일 11:45

SkillsMP 카탈로그 업데이트: 2026년 11월 5일 19:57

수집된 skills: 360

GitHub 스타: 275
GitHub 포크: 3

AI 어시스턴트로 설치

이 프롬프트를 복사해 사용 중인 AI 어시스턴트에 입력하세요.

Follow https://skillsmp.com/skill-install/prompt.md to install Agent Skills from https://github.com/criptogus/agent-evolve-network.

저장소 README

README 원본 파일: README.md

README 원본 보기

README 전체 펼치기

<div align="center"> <img src="https://superagentskill.com/og.png" alt="Super Agent Skill" width="640" /> # Super Agent Skill ### **The signed skill registry for AI agents.** > **Hundreds of packages, with tooling for adversarial evaluation, Ed25519 signing and evidence-based Trust Scores. Check each version’s report: publication does not mean verification.** > (Live count on [superagentskill.com/marketplace](https://superagentskill.com/marketplace).) **Skills, playbooks, souls (drop-in expert personas) and guardrails — one MCP endpoint or `npx super-agent install` away from any AI agent.** Think of it as a **library of verifiable expertise** for AI: pick a skill, beam it in, and your agent suddenly knows OWASP, MEDDPICC, ECG triage, SOC 2 evidence collection, or whatever your domain demands. Built for teams that have to **stand behind** what their AI does — security, fintech, healthcare, SOC 2 and DevOps. Works with Claude, Cursor, ChatGPT, Continue, Cline. No fine-tuning. No vendor lock-in. [![Website](https://img.shields.io/badge/🌐_superagentskill.com-Visit-0A66FF?style=for-the-badge)](https://superagentskill.com) [![Marketplace](https://img.shields.io/badge/🛒_Browse_the_registry-Free-8b5cf6?style=for-the-badge)](https://superagentskill.com/marketplace) [![MCP](https://img.shields.io/badge/🔌_MCP_endpoint-Live-22c55e?style=for-the-badge)](https://superagentskill.com/api/mcp) [![License: Apache 2.0](https://img.shields.io/badge/code-Apache_2.0-000?style=flat-square)](LICENSE) [![License: CC BY-SA 4.0](https://img.shields.io/badge/content-CC_BY--SA_4.0-000?style=flat-square)](LICENSES/CONTENT-CC-BY-SA-4.0.txt) [![PRs Welcome](https://img.shields.io/badge/PRs-welcome-ff4f8b?style=flat-square)](CONTRIBUTING.md) [![Discord](https://img.shields.io/badge/Discord-join-5865F2?style=flat-square)](https://superagentskill.com/community) [![X / Twitter](https://img.shields.io/badge/X-@gustavocaetano-000000?style=flat-square)](https://x.com/gustavocaetano) [**🌐 Website**](https://superagentskill.com) · [**🛒 Marketplace**](https://superagentskill.com/marketplace) · [**📚 Docs**](https://superagentskill.com/docs) · [**🔌 MCP**](https://superagentskill.com/api/mcp) · [**💬 Discord**](https://superagentskill.com/community) · [**🐦 X / Twitter**](https://x.com/gustavocaetano) </div> --- ## 🧬 Why this exists Models are commodities. **Expertise is the moat.** Today there are abundant models, agents, MCPs and workflows — but almost no one is dominating the **distribution of intelligence**. That layer compounds. Super Agent Skill turns any model into a domain expert in seconds — plug in, install a skill, and, well… *"I know kung fu."* Open source by default, community-curated, free to download, and continuously evolved by an automated quality loop (SkillForge). - **Free tier** — the open registry. Hundreds of packages (live count on the site). Fork, remix, install, redistribute. - **Premium tier** — [superagentskill.com](https://superagentskill.com) curates battle-tested elite stacks, enterprise playbooks and private packs for teams that want depth, not just breadth. The open network feeds the premium layer. The premium layer funds the open network. **That's the flywheel.** --- ## 📦 What's inside (today) > **The registry is live and growing every week. All packages are free to download, fork and run.** > The **In this repo** column is the open seed you can `git clone` today; the **Hosted registry** column is the full, continuously-evolved library on [superagentskill.com](https://superagentskill.com/marketplace). | Type | In this repo (open seed) | Hosted registry | Examples | |---|---|---|---| | 🛠 **Skills** | **104** | **370+** | code review, OWASP audit, SQL translator, OSINT, ECG triage, MEDDPICC discovery | | 📘 **Playbooks** | **2** | **30+** | bug triage → fix → PR, content pipeline, incident response | | 🎭 **Souls** | **6** | **50+** | pragmatic SRE, empathetic support, senior product designer | | 🛡 **Guardrails** | **2** | **50+** | no-PII, no-medical-advice, GDPR, brand safety | ➡️ **[Browse the full marketplace at superagentskill.com →](https://superagentskill.com/marketplace)** --- ## ⚡ 60-second quickstart ### Option A — Plug it into your agent (recommended) Point any MCP-compatible client at: ``` https://superagentskill.com/api/mcp ``` That's it. Your agent now has `discover` / `install` / `evaluate` tools and **live access to every skill in the registry** — including patches shipped overnight by SkillForge. ### Option B — Download a skill as a file (no account, free forever) ```bash # Grab a single skill curl -O https://raw.githubusercontent.com/criptogus/agent-evolve-network/main/content/skills/code-reviewer.yaml # Or clone the whole library git clone https://github.com/criptogus/agent-evolve-network.git ``` Drop the YAML into your agent's system prompt. Done. ### Option C — Need a skill that doesn't exist yet? Go to **[superagentskill.com](https://superagentskill.com)**, describe what your agent should do (e.g. *"triage cardiology patients using the GRACE score"*) and the **Forge** will research, draft, adversarially test and publish a production-grade package — usually in under 5 minutes. ### Option D — Run the whole platform locally ```bash bun install && bun run dev ``` --- ## 🧠 Why Super Agent Skill? **The problem:** anyone can write a YAML prompt and call it an "AI skill." The hard part is convincing your security team it won't leak PII, accept a jailbreak, or hallucinate a compliance disclosure in front of a customer. **What we ship that other registries don't:** - 🛡 **Proprietary adversarial harness.** The evaluation tooling can benchmark skills against attacks — prompt injection, jailbreaks, exfiltration, blast-radius, policy bypass, PII/PHI leakage. Cases are curated per vertical (OWASP LLM, FINRA, HIPAA Safe Harbor, PCI-DSS, SRE). - 🔏 **Ed25519-signed releases.** Published versions can carry cryptographic signatures; check the specific release for a signature. Verify offline with `npm run release:verify` — exactly what air-gapped enterprise customers need. - 🧾 **Signed per-package trust attestations.** `npm run trust:attest` binds a package's exact content hash + version to its adversarial result and signs it with the release key. Your security team runs `npm run trust:verify -- --attestation pkg.trust.json --pkg pkg.yaml --pubkey pub.pem` and confirms offline that *this exact file* passed *N* adversarial cases — no need to trust the hosted Trust Score endpoint. - 📊 **Public, verifiable Trust Score.** A transparent weighted formula over adversarial robustness, real-world success rate, signed releases and package age. Embed the badge in your README; reviewers see the same number you do. - 🏛 **Verticalized Souls.** Fintech compliance officer (FINRA / Reg E / PCI-DSS), HIPAA-aware clinical liaison, SOC 2 auditor (TSC-mapped), Kubernetes SRE (blast-radius-first). Souls that cite the rule before the recommendation. - 🔌 **One MCP endpoint, every IDE.** Claude, Cursor, ChatGPT, Continue, Cline — same URL, same `npx super-agent install <slug>` command. No SDK, no glue code. - 🛠 **SkillForge evolution loop.** Re-scores published skills against the live adversarial suite + telemetry; proposes patches automatically when robustness drops. > *Other registries hand you prompts. We hand you proof.* ## 🔍 Verify it yourself Don't take our word for it: ```bash # Verify a package's signed trust attestation offline npm run trust:verify -- --attestation pkg.trust.json --pkg pkg.yaml --pubkey pub.pem ``` Every package also has a public Trust page at `superagentskill.com/marketplace/trust/<slug>` showing its adversarial pass rate, signature status and Trust Score breakdown — the same numbers your reviewers will see. ## 👥 Who it's for | If you're… | What you get | | --- | --- | | 🚢 **Eng leadership shipping AI features** | Trust Score on the deploy checklist instead of a Notion doc. Real-world success rate per skill. | | 🔒 **Security / compliance officer** | Offline-verifiable signatures. Adversarial pass rate broken down by attack class. Upload-time injection-guard audit log. | | 🏥 **Fintech / healthcare / SOC 2 team** | Souls that cite FINRA 2210, Reg E §1005.7, HIPAA Safe Harbor, TSC CC6.x. PII/PHI guardrails redact PAN/CVV/MRNs before output. | | 💻 **Solo builder using Claude, Cursor or Continue** | `npx super-agent install <slug>` drops the right file into every IDE you use. No login for public skills. | | ✍️ **Skill author** | Premium revenue share, 12-month referral window, perpetual lineage cut on every fork, open bounties from companies. | --- ## 🧱 The four building blocks | | What it is | Use it when… | | --- | --- | --- | | 🛠 **Skills** | A focused capability with system prompt, rules, examples and a runnable input → output contract | You want the agent to *do one thing* very well (review code, write SQL, summarize a meeting). | | 📘 **Playbooks** | A multi-step decision graph that orchestrates skills + tools | The task has a recurring shape (triage bug → reproduce → propose fix → open PR). | | 🎭 **Souls** | A persona / value bundle — voice, taste, defaults — that wraps any agent. **Always free.** Souls inspired by public figures are fictional AI personas — not the real person ([disclaimer](SOULS_DISCLAIMER.md)). | You need a consistent brand, tone or worldview across all tasks. | | 🛡 **Guardrails** | Refusal, safety and compliance policies enforced before output | You ship to real users and need hard "must / must not" rules. | Each one is a single self-contained YAML file. **Read it, fork it, ship it.** ``` content/ ├── skills/ → Single capabilities (370+ on superagentskill.com) ├── playbooks/ → Multi-step workflows (30+ on superagentskill.com) ├── souls/ → Personas and values (50+ on superagentskill.com) ├── guardrails/ → Safety and compliance (50+ on superagentskill.com) └── schemas/ → JSON Schemas (every package validates against these) ``` > 💡 This GitHub repo is the **open seed** — the canonical, public-domain core. The full library (hundreds of packages, live count on the site) lives at [superagentskill.com/marketplace](https://superagentskill.com/marketplace), where SkillForge keeps it sharp. --- ## 🎯 Need to enrich a specific agent? Whatever your agent does, there's almost certainly a package for it — or you can have one in minutes: | If your agent is a… | Start here | |---|---| | **Coding assistant** | [code-reviewer](content/skills/code-reviewer.yaml), [owasp-code-audit](content/skills/owasp-code-audit.yaml), [dependency-vuln-auditor](content/skills/dependency-vuln-auditor.yaml) | | **Security / red team** | [recon-attack-surface](content/skills/recon-attack-surface.yaml), [prompt-injection-tester](content/skills/prompt-injection-tester.yaml), [disk-image-forensics](content/skills/disk-image-forensics.yaml) | | **SRE / DevOps** | [incident-response-triage](content/skills/incident-response-triage.yaml), [cloud-misconfig-auditor](content/skills/cloud-misconfig-auditor.yaml), [pragmatic-sre](content/souls/pragmatic-sre.yaml) | | **Researcher / writer** | [web-research](content/skills/web-research.yaml), [long-form-writer](content/skills/long-form-writer.yaml), [osint-investigator](content/skills/osint-investigator.yaml) | | **Customer support** | [empathetic-support](content/souls/empathetic-support.yaml), [no-pii-in-output](content/guardrails/no-pii-in-output.yaml) | | **Anything else** | 👉 **[Search the marketplace →](https://superagentskill.com/marketplace)** or **[ask the Forge to build it](https://superagentskill.com/forge)** | --- ## 🚢 Ship your first skill in 4 steps ```bash # 1. Fork & clone git clone git@github.com:<you>/agent-evolve-network.git cd agent-evolve-network && bun install # 2. Start from a template cp content/skills/_template.yaml content/skills/my-killer-skill.yaml # 3. Validate locally (same script CI runs) bun run validate:content # 4. Open a PR — a maintainer reviews within 72h ``` A merged skill is automatically imported into the hosted registry, where **SkillForge starts evolving it against adversarial benchmarks** and serving the best variant through MCP to every connected agent on day one. ➡️ Full quality bar in [**CONTRIBUTING.md**](CONTRIBUTING.md). --- ## 💎 Why contribute? Your work isn't just merged — it's *amplified*. - 🌐 **Reach.** Every accepted package is served live through MCP to thousands of agents. - 🏆 **Reputation.** Public **Trust Score**, model compatibility matrix and leaderboard rankings on the [marketplace](https://superagentskill.com/marketplace/rankings). - 💸 **Revenue share.** Premium packages earn 80–85% of sales + recurring usage bonuses. - 🤖 **AI co-author.** SkillForge proposes patches when your skill drifts — you stay in control, accept or reject. - 🪪 **CVE-style credit.** Robustness findings (e.g. `SAS-2026-0001`) are publicly attributed. > **First-time contributor?** Look for [`good first issue`](https://github.com/criptogus/agent-evolve-network/labels/good%20first%20issue) and [`help wanted`](https://github.com/criptogus/agent-evolve-network/labels/help%20wanted) — most are mergeable in one afternoon. --- ## 🆚 What makes this different | | Super Agent Skill | A prompt library on GitHub | A walled-garden GPT store | | --- | :---: | :---: | :---: | | Open content (CC BY-SA) | ✅ | ✅ | ❌ | | Schema-validated quality bar | ✅ | ❌ | ⚠️ opaque | | Works in *any* MCP agent | ✅ | ⚠️ manual paste | ❌ single vendor | | Continuous adversarial evolution | ✅ | ❌ | ❌ | | Public trust + compatibility scores | ✅ | ❌ | ❌ | | Revenue share for authors | ✅ | ❌ | ⚠️ limited | | One-line install via MCP | ✅ | ❌ | ❌ | --- ## 🗺 Roadmap - [x] MCP server with discover / install / evaluate / report-execution tools - [x] SkillForge evolution loop (eval → patch → re-score) - [x] Public **Trust Score**, model compatibility matrix, drift detection - [x] Hundreds of packages live in the hosted registry (live count on the site) - [x] CLI installation and connection commands (see `cli/`) - [x] CLI authoring preflight and immutable private draft upload (public approval remains gated) - [x] Cryptographic signing and verification tooling - [ ] Expand hosted signature and evaluation coverage with version-specific receipts - [x] Versioned registry bundles with signed archives, signed manifest and offline verification ([first release](https://github.com/criptogus/agent-evolve-network/releases/tag/content-v0.1.0)) - [x] Community collection creation, editing, ordering, following and public browsing (deployment pending) See [implementation and activation status](docs/ROADMAP-STATUS.md) before treating a checkbox as production evidence. Track everything on the [public board](https://github.com/criptogus/agent-evolve-network/projects). --- ## 👥 Community - 💬 **Discord** — [join here](https://superagentskill.com/community) for help, design reviews, and weekly skill-jam events. - 🐦 **X / Twitter** — [@superagentskill](https://twitter.com/superagentskill) for releases and benchmarks. - 📰 **Changelog** — every accepted package is announced in [GitHub Releases](https://github.com/criptogus/agent-evolve-network/releases). - 🛡 **Security** — please disclose responsibly via [SECURITY.md](SECURITY.md). --- ## 📜 License - **Code** (this repo, outside `content/`): [Apache License 2.0](LICENSE) - **Content** (everything in `content/`): [Creative Commons Attribution-ShareAlike 4.0](LICENSES/CONTENT-CC-BY-SA-4.0.txt) By contributing you agree your work is licensed under the same terms — and that it gets to make millions of agent conversations a little smarter every day. --- <div align="center"> ### 🚀 Ready to make your agent a specialist? **[→ Browse the free registry at superagentskill.com](https://superagentskill.com/marketplace)** **Built by makers, for makers. Star the repo ⭐ if you believe agents should belong to everyone.** </div>

저장소 README를 불러오는 중…

이 저장소의 skills

시장조사 분석가·마케팅 전문가 소프트웨어 개발자 데이터 과학자 정보 보안 분석가

직업 카테고리 18개 · 37% 분류됨

정보 보안 분석가

Inventories project dependencies and runtimes across ecosystems and reports known CVEs, supply-chain risks and a prioritized upgrade plan. Use when the user asks for dependency vulnerability auditor work, or mentions dependency, vuln, auditor.

원문 언어: 영어

2026-08-10

수집된 skill 360개 중 40개를 표시합니다.