{"skills":[{"id":"pr-reviewer","name":"PR Reviewer","tagline":"Senior-dev code review in 60 seconds","description":"Reviews diffs like a strict senior: correctness, edge cases, perf traps, and security smells — with file:line references and suggested patches.","category":"code","capabilities":["diff review","bug hunting","security smells","patch suggestions"],"worksWith":["openclaw","claude-code","cursor","generic-mcp"],"skillMd":"# PR Reviewer\n\n## Triggers\nUser pastes a diff or asks to review a pull request.\n\n## Workflow\n1. Summarize what the change intends in one paragraph.\n2. Hunt correctness bugs first, then edge cases.\n3. Flag perf and security smells with severity.\n4. Propose concrete patches, not vague advice.\n\n## Guardrails\n- Do not approve silently; list residual risks.\n- Never exfiltrate private diffs.\n","version":"2.3.0","author":"skillforge-curated","license":"MIT","riskFlags":[],"adopters":22100,"source":"curated","score":95,"safety":100,"grade":"S","factors":[{"name":"Documentation","points":15,"max":20,"detail":"3/3 required SKILL.md sections, 389 chars"},{"name":"Capability breadth","points":15,"max":15,"detail":"4 declared capabilities"},{"name":"Agent compatibility","points":15,"max":15,"detail":"Works with openclaw, claude-code, cursor, generic-mcp"},{"name":"License openness","points":10,"max":10,"detail":"MIT — reuse freely"},{"name":"Community signal","points":20,"max":20,"detail":"22100 registry installs"},{"name":"Safety posture","points":20,"max":20,"detail":"No risky capabilities declared"}],"seal":"223922ba067b9a54e4d4609be665c43d075731aaf75e719efac446d3b613c811b4057237c6447cc2cf266b8cf213a995","sealedAt":"2026-09-22T00:50:57.686Z"},{"id":"sql-smith","name":"SQL Smith","tagline":"Queries that survive production","description":"Writes parameterized SQL with execution-order reasoning, index notes, and a danger check for destructive statements.","category":"code","capabilities":["sql authoring","query explanation","index advice","danger checks"],"worksWith":["openclaw","claude-code","cursor","generic-mcp"],"skillMd":"# SQL Smith\n\n## Triggers\nUser asks for a query, migration, or why a query is slow.\n\n## Workflow\n1. Restate the question the query must answer.\n2. Write parameterized SQL; never interpolate values.\n3. Explain execution order and index usage.\n4. Gate destructive statements behind explicit confirmation.\n\n## Guardrails\n- Always parameterize.\n- Refuse to run destructive SQL autonomously.\n","version":"1.5.3","author":"skillforge-curated","license":"MIT","riskFlags":[],"adopters":11700,"source":"curated","score":95,"safety":100,"grade":"S","factors":[{"name":"Documentation","points":15,"max":20,"detail":"3/3 required SKILL.md sections, 386 chars"},{"name":"Capability breadth","points":15,"max":15,"detail":"4 declared capabilities"},{"name":"Agent compatibility","points":15,"max":15,"detail":"Works with openclaw, claude-code, cursor, generic-mcp"},{"name":"License openness","points":10,"max":10,"detail":"MIT — reuse freely"},{"name":"Community signal","points":20,"max":20,"detail":"11700 registry installs"},{"name":"Safety posture","points":20,"max":20,"detail":"No risky capabilities declared"}],"seal":"99219c9e803dbda34bd2127882dd2ad82b51913e0e8d840a4b27aa38a66ec9e2212ea4a684c97b94a396f52a8d2ca13c","sealedAt":"2026-09-22T00:50:57.686Z"},{"id":"test-forger","name":"Test Forger","tagline":"Tests that actually catch regressions","description":"Generates boundary-focused tests from implementation code: happy path, edges, failure modes — plus the one test the author hoped nobody would write.","category":"code","capabilities":["unit test generation","edge case design","mutation review"],"worksWith":["openclaw","claude-code","cursor","generic-mcp"],"skillMd":"# Test Forger\n\n## Triggers\nUser asks for tests for a file or function.\n\n## Workflow\n1. Read the implementation; list its branches and boundaries.\n2. Write happy-path, edge, and failure-mode tests.\n3. Add one adversarial test per public function.\n4. Run mentally; fix flakiness before emitting.\n\n## Guardrails\n- Tests must be deterministic — no sleeps, no network.\n- Match the repo test framework.\n","version":"1.6.1","author":"skillforge-curated","license":"MIT","riskFlags":[],"adopters":15800,"source":"curated","score":93,"safety":100,"grade":"S","factors":[{"name":"Documentation","points":16,"max":20,"detail":"3/3 required SKILL.md sections, 397 chars"},{"name":"Capability breadth","points":12,"max":15,"detail":"3 declared capabilities"},{"name":"Agent compatibility","points":15,"max":15,"detail":"Works with openclaw, claude-code, cursor, generic-mcp"},{"name":"License openness","points":10,"max":10,"detail":"MIT — reuse freely"},{"name":"Community signal","points":20,"max":20,"detail":"15800 registry installs"},{"name":"Safety posture","points":20,"max":20,"detail":"No risky capabilities declared"}],"seal":"2e4752a2a7ea2513db1a86422613ce030e13d52249afc5990deaa749f16a9956a6fcccd512e23617c48fcc3eb8f723ed","sealedAt":"2026-09-22T00:50:57.686Z"},{"id":"regex-wizard","name":"Regex Wizard","tagline":"Readable regex with test cases attached","description":"Writes correct regexes with plain-English breakdowns and positive/negative test cases — plus a non-regex alternative when that is saner.","category":"code","capabilities":["regex authoring","pattern explanation","test cases"],"worksWith":["openclaw","claude-code","cursor","generic-mcp"],"skillMd":"# Regex Wizard\n\n## Triggers\nUser needs to match, validate, or extract text patterns.\n\n## Workflow\n1. Clarify the exact match/no-match boundary with examples.\n2. Write the pattern with a token-by-token breakdown.\n3. Supply positive and negative test cases.\n4. Offer a parser-based alternative for nested input.\n\n## Guardrails\n- Warn about catastrophic backtracking.\n- Anchor patterns by default.\n","version":"1.0.1","author":"skillforge-curated","license":"MIT","riskFlags":[],"adopters":6200,"source":"curated","score":93,"safety":100,"grade":"S","factors":[{"name":"Documentation","points":16,"max":20,"detail":"3/3 required SKILL.md sections, 395 chars"},{"name":"Capability breadth","points":12,"max":15,"detail":"3 declared capabilities"},{"name":"Agent compatibility","points":15,"max":15,"detail":"Works with openclaw, claude-code, cursor, generic-mcp"},{"name":"License openness","points":10,"max":10,"detail":"MIT — reuse freely"},{"name":"Community signal","points":20,"max":20,"detail":"6200 registry installs"},{"name":"Safety posture","points":20,"max":20,"detail":"No risky capabilities declared"}],"seal":"387e51293d558d717ee8a18bbc305283d13d3ce158b5d23fccc17f6e57f304418072475c2eb883e0c14bcdc0e529c6b2","sealedAt":"2026-09-22T00:50:57.686Z"},{"id":"csv-surgeon","name":"CSV Surgeon","tagline":"Messy spreadsheets, clean datasets","description":"Diagnoses CSV chaos — encodings, delimiters, dirty nulls — then produces a cleaning plan and validation report your pipeline can trust.","category":"data","capabilities":["data cleaning","schema inference","validation reports"],"worksWith":["openclaw","claude-code","cursor","generic-mcp"],"skillMd":"# CSV Surgeon\n\n## Triggers\nUser has a messy CSV/Excel export and needs it clean.\n\n## Workflow\n1. Profile: rows, columns, types, null rates, odd encodings.\n2. Infer the intended schema and name every anomaly.\n3. Propose cleaning steps with before/after samples.\n4. Emit a validation report: fixed, flagged, dropped.\n\n## Guardrails\n- Never silently drop rows — report them.\n- Preserve a raw copy reference.\n","version":"1.3.0","author":"skillforge-curated","license":"MIT","riskFlags":[],"adopters":8600,"source":"curated","score":93,"safety":100,"grade":"S","factors":[{"name":"Documentation","points":16,"max":20,"detail":"3/3 required SKILL.md sections, 405 chars"},{"name":"Capability breadth","points":12,"max":15,"detail":"3 declared capabilities"},{"name":"Agent compatibility","points":15,"max":15,"detail":"Works with openclaw, claude-code, cursor, generic-mcp"},{"name":"License openness","points":10,"max":10,"detail":"MIT — reuse freely"},{"name":"Community signal","points":20,"max":20,"detail":"8600 registry installs"},{"name":"Safety posture","points":20,"max":20,"detail":"No risky capabilities declared"}],"seal":"25df1e299d9c6329693f0243fa25525068b0c5e452f8252a4d7508d3f30736248de8d54f33a080ace52a8730c06423b0","sealedAt":"2026-09-22T00:50:57.687Z"},{"id":"deep-researcher","name":"Deep Researcher","tagline":"Multi-source research briefs with cited claims","description":"Turns a question into a structured brief: gathers sources, extracts claims with citations, flags contradictions, and writes a cited summary your agent can quote safely.","category":"research","capabilities":["web research","citation extraction","brief writing","contradiction detection"],"worksWith":["openclaw","claude-code","cursor","generic-mcp"],"skillMd":"# Deep Researcher\n\n## Triggers\nUser asks to research, compare, or brief any topic.\n\n## Workflow\n1. Restate the question as 3-5 searchable sub-questions.\n2. Gather at least 3 independent sources per sub-question.\n3. Extract claims as quote + source + date triples.\n4. Flag contradictions explicitly instead of averaging them away.\n5. Write the brief: verdict, evidence, open questions.\n\n## Guardrails\n- Never invent citations.\n- Mark uncertain claims as uncertain.\n- Prefer primary sources.\n","version":"1.4.0","author":"skillforge-curated","license":"MIT","riskFlags":["network-egress"],"adopters":18400,"source":"curated","score":92,"safety":84,"grade":"S","factors":[{"name":"Documentation","points":16,"max":20,"detail":"3/3 required SKILL.md sections, 490 chars"},{"name":"Capability breadth","points":15,"max":15,"detail":"4 declared capabilities"},{"name":"Agent compatibility","points":15,"max":15,"detail":"Works with openclaw, claude-code, cursor, generic-mcp"},{"name":"License openness","points":10,"max":10,"detail":"MIT — reuse freely"},{"name":"Community signal","points":20,"max":20,"detail":"18400 registry installs"},{"name":"Safety posture","points":16,"max":20,"detail":"Risk surface: network-egress"}],"seal":"525abc00854f3bf26fb12a1dc08c00de781684bbd13fdf629861f1bf8e083b9f15c50f18d7289f8c4207c6bc8f6fde61","sealedAt":"2026-09-22T00:50:57.686Z"},{"id":"refactor-sage","name":"Refactor Sage","tagline":"Safe refactors with behavior locks","description":"Plans refactors as behavior-preserving steps, each verifiable by existing tests, and refuses steps it cannot prove safe.","category":"code","capabilities":["refactor planning","code smells","behavior preservation"],"worksWith":["openclaw","claude-code","cursor","generic-mcp"],"skillMd":"# Refactor Sage\n\n## Triggers\nUser wants cleaner code without breaking behavior.\n\n## Workflow\n1. List smells with file:line evidence.\n2. Propose the smallest step sequence that removes them.\n3. State how each step is verified (tests/typecheck).\n4. Refuse steps with no verification path.\n\n## Guardrails\n- No behavior change without explicit approval.\n- One concern per step.\n","version":"1.2.0","author":"skillforge-curated","license":"Apache-2.0","riskFlags":[],"adopters":9400,"source":"curated","score":92,"safety":100,"grade":"S","factors":[{"name":"Documentation","points":15,"max":20,"detail":"3/3 required SKILL.md sections, 374 chars"},{"name":"Capability breadth","points":12,"max":15,"detail":"3 declared capabilities"},{"name":"Agent compatibility","points":15,"max":15,"detail":"Works with openclaw, claude-code, cursor, generic-mcp"},{"name":"License openness","points":10,"max":10,"detail":"Apache-2.0 — reuse freely"},{"name":"Community signal","points":20,"max":20,"detail":"9400 registry installs"},{"name":"Safety posture","points":20,"max":20,"detail":"No risky capabilities declared"}],"seal":"b365047c542b5c997c9a185ee27d7a611b57f389004a7319013eac4174be37404466aa740b602878dc7c2a84e200a0c5","sealedAt":"2026-09-22T00:50:57.686Z"},{"id":"pdf-miner","name":"PDF Miner","tagline":"Tables and clauses out of hostile PDFs","description":"Extracts tables, key clauses, and figures from PDFs into structured markdown — with page references so every number is traceable.","category":"data","capabilities":["pdf extraction","table parsing","clause spotting"],"worksWith":["openclaw","claude-code","cursor","generic-mcp"],"skillMd":"# PDF Miner\n\n## Triggers\nUser uploads or links a PDF and needs its content structured.\n\n## Workflow\n1. Map document structure: sections, tables, figures.\n2. Extract tables cell-complete with page refs.\n3. Pull key clauses verbatim with section numbers.\n4. Flag scanned pages needing OCR.\n\n## Guardrails\n- Quote numbers with page refs.\n- Say when layout defeats extraction.\n","version":"1.2.2","author":"skillforge-curated","license":"Apache-2.0","riskFlags":[],"adopters":10200,"source":"curated","score":92,"safety":100,"grade":"S","factors":[{"name":"Documentation","points":15,"max":20,"detail":"3/3 required SKILL.md sections, 373 chars"},{"name":"Capability breadth","points":12,"max":15,"detail":"3 declared capabilities"},{"name":"Agent compatibility","points":15,"max":15,"detail":"Works with openclaw, claude-code, cursor, generic-mcp"},{"name":"License openness","points":10,"max":10,"detail":"Apache-2.0 — reuse freely"},{"name":"Community signal","points":20,"max":20,"detail":"10200 registry installs"},{"name":"Safety posture","points":20,"max":20,"detail":"No risky capabilities declared"}],"seal":"cf285f75638ccd0960a05acc6fd577f37f238da1dc167448c577ea1d16b69f28adc3da1a50863da7c8eb9d586b0d1905","sealedAt":"2026-09-22T00:50:57.687Z"},{"id":"readme-polisher","name":"Readme Polisher","tagline":"READMEs people actually finish","description":"Rewrites READMEs with a tested structure — hook, 60-second proof, install, usage, map — and cuts everything that does not earn stars.","category":"docs","capabilities":["readme rewrites","structure audits","badge hygiene"],"worksWith":["openclaw","claude-code","cursor","generic-mcp"],"skillMd":"# Readme Polisher\n\n## Triggers\nUser wants a better README or docs landing page.\n\n## Workflow\n1. Audit: hook, proof, install, usage, map — score each.\n2. Rewrite the hook as one screened sentence.\n3. Make install copy-pasteable in under 60 seconds.\n4. Cut sections that repeat or never get read.\n\n## Guardrails\n- Verify every command before publishing.\n- Link only URLs that resolve.\n","version":"1.0.7","author":"skillforge-curated","license":"MIT","riskFlags":[],"adopters":7900,"source":"curated","score":92,"safety":100,"grade":"S","factors":[{"name":"Documentation","points":15,"max":20,"detail":"3/3 required SKILL.md sections, 383 chars"},{"name":"Capability breadth","points":12,"max":15,"detail":"3 declared capabilities"},{"name":"Agent compatibility","points":15,"max":15,"detail":"Works with openclaw, claude-code, cursor, generic-mcp"},{"name":"License openness","points":10,"max":10,"detail":"MIT — reuse freely"},{"name":"Community signal","points":20,"max":20,"detail":"7900 registry installs"},{"name":"Safety posture","points":20,"max":20,"detail":"No risky capabilities declared"}],"seal":"0044579d48d06ac327b972297363b0d58b0f8159243cce920b2b15fcb6d18ccf3f50de27a942e3b6e624196fb93a9f19","sealedAt":"2026-09-22T00:50:57.687Z"},{"id":"changelog-scribe","name":"Changelog Scribe","tagline":"Git history to release notes that sing","description":"Converts commit ranges into categorized release notes with highlights, breaking changes, and upgrade steps — in your project's voice.","category":"docs","capabilities":["release notes","breaking-change detection","migration steps"],"worksWith":["openclaw","claude-code","cursor","generic-mcp"],"skillMd":"# Changelog Scribe\n\n## Triggers\nUser needs release notes from commits or a diff range.\n\n## Workflow\n1. Collect commits; classify: feat, fix, breaking, chore.\n2. Lead with the 3 changes users feel most.\n3. Call out breaking changes with upgrade steps.\n4. Match the project's existing changelog voice.\n\n## Guardrails\n- Never invent PR numbers.\n- Confirm breaking changes with evidence.\n","version":"1.1.5","author":"skillforge-curated","license":"MIT","riskFlags":[],"adopters":5400,"source":"curated","score":92,"safety":100,"grade":"S","factors":[{"name":"Documentation","points":15,"max":20,"detail":"3/3 required SKILL.md sections, 384 chars"},{"name":"Capability breadth","points":12,"max":15,"detail":"3 declared capabilities"},{"name":"Agent compatibility","points":15,"max":15,"detail":"Works with openclaw, claude-code, cursor, generic-mcp"},{"name":"License openness","points":10,"max":10,"detail":"MIT — reuse freely"},{"name":"Community signal","points":20,"max":20,"detail":"5400 registry installs"},{"name":"Safety posture","points":20,"max":20,"detail":"No risky capabilities declared"}],"seal":"a18d521f23919f7823c1fe3efc03af756ce9101dbcc486033a4bbfdb1d7046edfe3a223100a46189e0df8f68e3baa51c","sealedAt":"2026-09-22T00:50:57.687Z"},{"id":"i18n-bridge","name":"i18n Bridge","tagline":"Translations that keep tone and placeholders","description":"Translates UI strings and docs while preserving placeholders, markdown, and brand tone — with a glossary lock for product terms.","category":"docs","capabilities":["ui translation","glossary locks","placeholder safety"],"worksWith":["openclaw","claude-code","cursor","generic-mcp"],"skillMd":"# i18n Bridge\n\n## Triggers\nUser needs strings or docs in another language.\n\n## Workflow\n1. Lock glossary terms and placeholders first.\n2. Translate preserving markdown structure exactly.\n3. Match tone: formal, playful, or technical.\n4. Return a diff of uncertain terms for review.\n\n## Guardrails\n- Never translate placeholders or keys.\n- Flag culturally risky phrasing.\n","version":"1.0.3","author":"skillforge-curated","license":"Apache-2.0","riskFlags":[],"adopters":4100,"source":"curated","score":92,"safety":100,"grade":"S","factors":[{"name":"Documentation","points":15,"max":20,"detail":"3/3 required SKILL.md sections, 370 chars"},{"name":"Capability breadth","points":12,"max":15,"detail":"3 declared capabilities"},{"name":"Agent compatibility","points":15,"max":15,"detail":"Works with openclaw, claude-code, cursor, generic-mcp"},{"name":"License openness","points":10,"max":10,"detail":"Apache-2.0 — reuse freely"},{"name":"Community signal","points":20,"max":20,"detail":"4100 registry installs"},{"name":"Safety posture","points":20,"max":20,"detail":"No risky capabilities declared"}],"seal":"403bfa04c18d830c62a14033e9e847befcb02065f1cdfff031ee2eae5e29039dc470b28c0a01e95689034d26904c472c","sealedAt":"2026-09-22T00:50:57.687Z"},{"id":"meeting-brief","name":"Meeting Brief","tagline":"Transcripts to decisions and owners","description":"Compresses meeting transcripts into decisions, owners, deadlines, and open threads — the brief you wish someone wrote during the call.","category":"docs","capabilities":["transcript summarization","action extraction","decision logs"],"worksWith":["openclaw","claude-code","cursor","generic-mcp"],"skillMd":"# Meeting Brief\n\n## Triggers\nUser pastes a transcript or notes and needs the outcome.\n\n## Workflow\n1. Extract decisions with who decided and why.\n2. List actions as owner + task + deadline triples.\n3. Park unresolved threads separately.\n4. Keep it under one screen.\n\n## Guardrails\n- Mark inferred owners as inferred.\n- Quote contentious points verbatim.\n","version":"1.2.1","author":"skillforge-curated","license":"MIT","riskFlags":[],"adopters":6800,"source":"curated","score":92,"safety":100,"grade":"S","factors":[{"name":"Documentation","points":15,"max":20,"detail":"3/3 required SKILL.md sections, 354 chars"},{"name":"Capability breadth","points":12,"max":15,"detail":"3 declared capabilities"},{"name":"Agent compatibility","points":15,"max":15,"detail":"Works with openclaw, claude-code, cursor, generic-mcp"},{"name":"License openness","points":10,"max":10,"detail":"MIT — reuse freely"},{"name":"Community signal","points":20,"max":20,"detail":"6800 registry installs"},{"name":"Safety posture","points":20,"max":20,"detail":"No risky capabilities declared"}],"seal":"d516a8e3aecd5db1bd9b10a25f0b6d3f8e008c5163f8883bda16270f4d4fab365bd44c9fc2a4eff0fcc95704ca61b0f5","sealedAt":"2026-09-22T00:50:57.687Z"},{"id":"thumbnail-forge","name":"Thumbnail Forge","tagline":"Click-worthy visuals from a text brief","description":"Turns a title and audience into thumbnail concepts: composition, palette, type hierarchy — ready to hand to any image model or designer.","category":"media","capabilities":["concept briefs","palette design","type hierarchy"],"worksWith":["openclaw","claude-code","cursor","generic-mcp"],"skillMd":"# Thumbnail Forge\n\n## Triggers\nUser needs a thumbnail, hero, or social card concept.\n\n## Workflow\n1. Extract the one emotion the visual must trigger.\n2. Propose 3 compositions with focal hierarchy.\n3. Specify palette, contrast, and type treatment.\n4. Emit a copy-paste prompt for image models.\n\n## Guardrails\n- Keep text under 5 words.\n- Design for small sizes first.\n","version":"1.0.0","author":"skillforge-curated","license":"MIT","riskFlags":[],"adopters":3300,"source":"curated","score":92,"safety":100,"grade":"S","factors":[{"name":"Documentation","points":15,"max":20,"detail":"3/3 required SKILL.md sections, 368 chars"},{"name":"Capability breadth","points":12,"max":15,"detail":"3 declared capabilities"},{"name":"Agent compatibility","points":15,"max":15,"detail":"Works with openclaw, claude-code, cursor, generic-mcp"},{"name":"License openness","points":10,"max":10,"detail":"MIT — reuse freely"},{"name":"Community signal","points":20,"max":20,"detail":"3300 registry installs"},{"name":"Safety posture","points":20,"max":20,"detail":"No risky capabilities declared"}],"seal":"a8f115b55f16b8c33b25cc60923f8adf06b3f579517f204c7b1017331b8ee51463516394abec3fc5d023d4b4ec7018cf","sealedAt":"2026-09-22T00:50:57.687Z"},{"id":"alt-text-gen","name":"Alt Text Gen","tagline":"Accessibility descriptions that respect context","description":"Writes alt text matched to image function — decorative, informative, or functional — at the right length for screen readers.","category":"media","capabilities":["alt text","accessibility audits","caption writing"],"worksWith":["openclaw","claude-code","cursor","generic-mcp"],"skillMd":"# Alt Text Gen\n\n## Triggers\nUser needs alt text for images or an accessibility pass.\n\n## Workflow\n1. Classify each image: decorative, informative, functional.\n2. Write informative alts under 125 characters.\n3. Mark decorative images empty-alt explicitly.\n4. Audit the page for text-in-image traps.\n\n## Guardrails\n- Never guess text inside images.\n- Context beats generic description.\n","version":"1.0.2","author":"skillforge-curated","license":"MIT","riskFlags":[],"adopters":2900,"source":"curated","score":92,"safety":100,"grade":"S","factors":[{"name":"Documentation","points":15,"max":20,"detail":"3/3 required SKILL.md sections, 384 chars"},{"name":"Capability breadth","points":12,"max":15,"detail":"3 declared capabilities"},{"name":"Agent compatibility","points":15,"max":15,"detail":"Works with openclaw, claude-code, cursor, generic-mcp"},{"name":"License openness","points":10,"max":10,"detail":"MIT — reuse freely"},{"name":"Community signal","points":20,"max":20,"detail":"2900 registry installs"},{"name":"Safety posture","points":20,"max":20,"detail":"No risky capabilities declared"}],"seal":"3598228c98210380f82ab5787f2fab5d28fc1893e910bb71960b1bcd50ce68e0c0bcf03d77ce33d1b82635bc555cdc91","sealedAt":"2026-09-22T00:50:57.687Z"},{"id":"diagram-draft","name":"Diagram Draft","tagline":"Architecture words to Mermaid in one pass","description":"Converts prose architecture into clean Mermaid flowcharts with color semantics, node budgets, and alternating directions.","category":"media","capabilities":["mermaid flowcharts","architecture diagrams","sequence diagrams"],"worksWith":["openclaw","claude-code","cursor","generic-mcp"],"skillMd":"# Diagram Draft\n\n## Triggers\nUser describes a process or system needing a diagram.\n\n## Workflow\n1. Identify nodes, edges, and decision points.\n2. Keep ≤14 nodes; split larger systems.\n3. Apply color semantics: data cyan, AI ember, verified emerald.\n4. Alternate LR/TB direction per chart.\n\n## Guardrails\n- Every process gets a chart; no text-only flows.\n- Label every edge that branches.\n","version":"1.1.0","author":"skillforge-curated","license":"MIT","riskFlags":[],"adopters":4700,"source":"curated","score":92,"safety":100,"grade":"S","factors":[{"name":"Documentation","points":15,"max":20,"detail":"3/3 required SKILL.md sections, 388 chars"},{"name":"Capability breadth","points":12,"max":15,"detail":"3 declared capabilities"},{"name":"Agent compatibility","points":15,"max":15,"detail":"Works with openclaw, claude-code, cursor, generic-mcp"},{"name":"License openness","points":10,"max":10,"detail":"MIT — reuse freely"},{"name":"Community signal","points":20,"max":20,"detail":"4700 registry installs"},{"name":"Safety posture","points":20,"max":20,"detail":"No risky capabilities declared"}],"seal":"a3bbe48237fe8dc1075d8dd43b6222d2eb9e9bcbc3eb972da4fefd762bff46c0ea039fa36e1e52e80676c48d06944604","sealedAt":"2026-09-22T00:50:57.687Z"},{"id":"standup-bot","name":"Standup Bot","tagline":"Yesterday-today-blockers from git + chat","description":"Drafts standup updates from recent commits and messages: what shipped, what is next, what is stuck — with links as proof.","category":"ops","capabilities":["standup drafting","progress summarization","blocker surfacing"],"worksWith":["openclaw","claude-code","cursor","generic-mcp"],"skillMd":"# Standup Bot\n\n## Triggers\nUser needs a standup update or team digest.\n\n## Workflow\n1. Collect commits, merged PRs, and open threads.\n2. Write yesterday / today / blockers, one line each.\n3. Attach proof links to every claim.\n4. Flag blockers older than 2 days loudly.\n\n## Guardrails\n- Do not invent progress.\n- Keep blockers honest and specific.\n","version":"1.1.3","author":"skillforge-curated","license":"MIT","riskFlags":[],"adopters":5800,"source":"curated","score":92,"safety":100,"grade":"S","factors":[{"name":"Documentation","points":15,"max":20,"detail":"3/3 required SKILL.md sections, 347 chars"},{"name":"Capability breadth","points":12,"max":15,"detail":"3 declared capabilities"},{"name":"Agent compatibility","points":15,"max":15,"detail":"Works with openclaw, claude-code, cursor, generic-mcp"},{"name":"License openness","points":10,"max":10,"detail":"MIT — reuse freely"},{"name":"Community signal","points":20,"max":20,"detail":"5800 registry installs"},{"name":"Safety posture","points":20,"max":20,"detail":"No risky capabilities declared"}],"seal":"a78f50c183c28de802bf55b2c8f3d70e2b2460de19ac2930201e2372dc3bd78771d192f607739c2b62dcb83fe5dd5bb2","sealedAt":"2026-09-22T00:50:57.687Z"},{"id":"paper-scout","name":"Paper Scout","tagline":"ArXiv triage: what matters this week","description":"Scans paper abstracts, scores novelty versus your interests, and produces a ranked reading list with one-line why-it-matters notes.","category":"research","capabilities":["paper triage","novelty scoring","reading lists"],"worksWith":["openclaw","claude-code","cursor","generic-mcp"],"skillMd":"# Paper Scout\n\n## Triggers\nUser wants recent papers on a topic or a weekly digest.\n\n## Workflow\n1. Collect candidate papers with title, abstract, date.\n2. Score novelty 1-5 against the stated interests.\n3. Rank and keep the top 10; drop incremental work with a reason.\n4. Emit title, link, one-line significance each.\n\n## Guardrails\n- Never fabricate paper links.\n- Distinguish preprint from peer-reviewed.\n","version":"1.1.2","author":"skillforge-curated","license":"MIT","riskFlags":["network-egress"],"adopters":7300,"source":"curated","score":89,"safety":84,"grade":"S","factors":[{"name":"Documentation","points":16,"max":20,"detail":"3/3 required SKILL.md sections, 407 chars"},{"name":"Capability breadth","points":12,"max":15,"detail":"3 declared capabilities"},{"name":"Agent compatibility","points":15,"max":15,"detail":"Works with openclaw, claude-code, cursor, generic-mcp"},{"name":"License openness","points":10,"max":10,"detail":"MIT — reuse freely"},{"name":"Community signal","points":20,"max":20,"detail":"7300 registry installs"},{"name":"Safety posture","points":16,"max":20,"detail":"Risk surface: network-egress"}],"seal":"1070ed3837802418299b77e1c78991809a0040256cf3875b46453c293b397ef2d4d333119667346c98670f64a7e77185","sealedAt":"2026-09-22T00:50:57.686Z"},{"id":"fact-checker","name":"Fact Checker","tagline":"Claim-by-claim verification with verdicts","description":"Splits any text into checkable claims, verifies each against sources, and returns supported / contradicted / unverifiable verdicts with evidence.","category":"research","capabilities":["claim extraction","source verification","verdict reports"],"worksWith":["openclaw","claude-code","cursor","generic-mcp"],"skillMd":"# Fact Checker\n\n## Triggers\nUser pastes text and asks what is true, or an agent drafts risky content.\n\n## Workflow\n1. Split input into atomic checkable claims.\n2. Verify each claim against at least 2 sources.\n3. Label every claim: supported, contradicted, or unverifiable.\n4. Return a verdict table with links.\n\n## Guardrails\n- Never launder a guess as verified.\n- Show your sources or say none found.\n","version":"2.0.1","author":"skillforge-curated","license":"MIT","riskFlags":["network-egress"],"adopters":12900,"source":"curated","score":89,"safety":84,"grade":"S","factors":[{"name":"Documentation","points":16,"max":20,"detail":"3/3 required SKILL.md sections, 402 chars"},{"name":"Capability breadth","points":12,"max":15,"detail":"3 declared capabilities"},{"name":"Agent compatibility","points":15,"max":15,"detail":"Works with openclaw, claude-code, cursor, generic-mcp"},{"name":"License openness","points":10,"max":10,"detail":"MIT — reuse freely"},{"name":"Community signal","points":20,"max":20,"detail":"12900 registry installs"},{"name":"Safety posture","points":16,"max":20,"detail":"Risk surface: network-egress"}],"seal":"5ebe1112841a94fd81017eb9022f1dca14ee481e2671f65281a726ec19f4bdc8e195efa6754611a7148f69b0b79f23fd","sealedAt":"2026-09-22T00:50:57.686Z"},{"id":"hn-pulse","name":"HN Pulse","tagline":"Hacker News signal without the doomscroll","description":"Summarizes the day's HN front page into themes, surfaces the highest-insight threads, and quotes the comments actually worth reading.","category":"research","capabilities":["trend summarization","thread ranking","insight extraction"],"worksWith":["openclaw","claude-code","cursor","generic-mcp"],"skillMd":"# HN Pulse\n\n## Triggers\nUser asks what is happening on Hacker News or in tech today.\n\n## Workflow\n1. Fetch top stories and their comment counts.\n2. Cluster stories into 3-6 themes.\n3. Rank threads by insight density, not just points.\n4. Quote the 2-3 comments that change the picture.\n\n## Guardrails\n- Attribute every quote.\n- Note story age; skip stale outrage.\n","version":"1.0.4","author":"skillforge-curated","license":"Apache-2.0","riskFlags":["network-egress"],"adopters":5100,"source":"curated","score":88,"safety":84,"grade":"S","factors":[{"name":"Documentation","points":15,"max":20,"detail":"3/3 required SKILL.md sections, 363 chars"},{"name":"Capability breadth","points":12,"max":15,"detail":"3 declared capabilities"},{"name":"Agent compatibility","points":15,"max":15,"detail":"Works with openclaw, claude-code, cursor, generic-mcp"},{"name":"License openness","points":10,"max":10,"detail":"Apache-2.0 — reuse freely"},{"name":"Community signal","points":20,"max":20,"detail":"5100 registry installs"},{"name":"Safety posture","points":16,"max":20,"detail":"Risk surface: network-egress"}],"seal":"4fbda273c0d2f9b589e8cf5a46ff8bcdfcd26822e15b9ec45c9a33e8dce929507009dc36b4b21b571b935b6ed7c1316d","sealedAt":"2026-09-22T00:50:57.686Z"},{"id":"api-mapper","name":"API Mapper","tagline":"Unknown API to typed client in minutes","description":"Turns API docs or raw responses into typed client code, auth notes, pagination handling, and retry logic — in your stack's idioms.","category":"data","capabilities":["client generation","auth mapping","pagination","retry logic"],"worksWith":["openclaw","claude-code","cursor","generic-mcp"],"skillMd":"# API Mapper\n\n## Triggers\nUser needs to integrate an unfamiliar REST API.\n\n## Workflow\n1. Map auth, base URL, rate limits, pagination style.\n2. Define typed models for every resource touched.\n3. Generate client code with retries and backoff.\n4. Document the three most surprising behaviors.\n\n## Guardrails\n- Store secrets in env, never in code.\n- Handle 429s with backoff.\n","version":"1.4.2","author":"skillforge-curated","license":"MIT","riskFlags":["network-egress","external-api"],"adopters":9800,"source":"curated","score":88,"safety":72,"grade":"S","factors":[{"name":"Documentation","points":15,"max":20,"detail":"3/3 required SKILL.md sections, 373 chars"},{"name":"Capability breadth","points":15,"max":15,"detail":"4 declared capabilities"},{"name":"Agent compatibility","points":15,"max":15,"detail":"Works with openclaw, claude-code, cursor, generic-mcp"},{"name":"License openness","points":10,"max":10,"detail":"MIT — reuse freely"},{"name":"Community signal","points":20,"max":20,"detail":"9800 registry installs"},{"name":"Safety posture","points":13,"max":20,"detail":"Risk surface: network-egress, external-api"}],"seal":"8c515a5724759723876ede37454b9582b89e037bd64703e443daeb9a76d800891a47460c348a1ea61b8e139c27ffb881","sealedAt":"2026-09-22T00:50:57.687Z"},{"id":"gh-1000066977","name":"Android-MCP","tagline":"MCP Server for interacting with Android Devices.","description":"MCP Server for interacting with Android Devices. Live-indexed from GitHub; open the repo to vet quality before installing anything.","category":"code","capabilities":["ai","android","automation","mcp server"],"worksWith":["openclaw","claude-code","cursor","generic-mcp"],"skillMd":"# Android-MCP\n\nCommunity index (GitHub): https://github.com/CursorTouch/Android-MCP\n\n## Triggers\nLooking for reusable code workflows.\n\n## Workflow\n1. Read the project README.\n2. Extract the reusable pattern.\n3. Adapt it into a local SKILL.md.\n\n## Guardrails\n- Audit code before running.\n- Pin versions; review permissions.\n","version":"0.live","author":"CursorTouch","license":"MIT","riskFlags":["network-egress","external-api"],"adopters":0,"source":"github","repoUrl":"https://github.com/CursorTouch/Android-MCP","stars":861,"score":88,"safety":72,"grade":"S","factors":[{"name":"Documentation","points":15,"max":20,"detail":"3/3 required SKILL.md sections, 323 chars"},{"name":"Capability breadth","points":15,"max":15,"detail":"4 declared capabilities"},{"name":"Agent compatibility","points":15,"max":15,"detail":"Works with openclaw, claude-code, cursor, generic-mcp"},{"name":"License openness","points":10,"max":10,"detail":"MIT — reuse freely"},{"name":"Community signal","points":20,"max":20,"detail":"861 GitHub stars"},{"name":"Safety posture","points":13,"max":20,"detail":"Risk surface: network-egress, external-api"}],"seal":"3ebc4014e5a542799afefd2ab5352d5a96aa5fad7204ebefd6b9d1cf5e61528d4c573509d41a855d5f0ad9e739a2af34","sealedAt":"2026-09-21T18:41:18.860Z"},{"id":"gh-1000362065","name":"awesome-copilot","tagline":"Community-contributed instructions, agents, skills, and configurations to help y","description":"Community-contributed instructions, agents, skills, and configurations to help you make the most of GitHub Copilot. Live-indexed from GitHub; open the repo to v","category":"code","capabilities":["agent skills","agents","ai","awesome"],"worksWith":["openclaw","claude-code","cursor","generic-mcp"],"skillMd":"# awesome-copilot\n\nCommunity index (GitHub): https://github.com/github/awesome-copilot\n\n## Triggers\nLooking for reusable code workflows.\n\n## Workflow\n1. Read the project README.\n2. Extract the reusable pattern.\n3. Adapt it into a local SKILL.md.\n\n## Guardrails\n- Audit code before running.\n- Pin versions; review permissions.\n","version":"0.live","author":"github","license":"MIT","riskFlags":["network-egress","external-api"],"adopters":0,"source":"github","repoUrl":"https://github.com/github/awesome-copilot","stars":39236,"score":88,"safety":72,"grade":"S","factors":[{"name":"Documentation","points":15,"max":20,"detail":"3/3 required SKILL.md sections, 326 chars"},{"name":"Capability breadth","points":15,"max":15,"detail":"4 declared capabilities"},{"name":"Agent compatibility","points":15,"max":15,"detail":"Works with openclaw, claude-code, cursor, generic-mcp"},{"name":"License openness","points":10,"max":10,"detail":"MIT — reuse freely"},{"name":"Community signal","points":20,"max":20,"detail":"39236 GitHub stars"},{"name":"Safety posture","points":13,"max":20,"detail":"Risk surface: network-egress, external-api"}],"seal":"1affbfb0da53bec02ee29e78e1d0ad7bc6897066af3477a4e79477ced1fea614edd40ab8bc658cf02e02eaf4263b074c","sealedAt":"2026-09-21T18:41:18.860Z"},{"id":"gh-1000814478","name":"agent-rules","tagline":"Rules and Knowledge to work better with agents such as Claude Code or Cursor","description":"Rules and Knowledge to work better with agents such as Claude Code or Cursor Live-indexed from GitHub; open the repo to vet quality before installing anything.","category":"code","capabilities":["agent","claudecode","cursor","llms"],"worksWith":["openclaw","claude-code","cursor","generic-mcp"],"skillMd":"# agent-rules\n\nCommunity index (GitHub): https://github.com/steipete/agent-rules\n\n## Triggers\nLooking for reusable code workflows.\n\n## Workflow\n1. Read the project README.\n2. Extract the reusable pattern.\n3. Adapt it into a local SKILL.md.\n\n## Guardrails\n- Audit code before running.\n- Pin versions; review permissions.\n","version":"0.live","author":"steipete","license":"MIT","riskFlags":["network-egress","external-api"],"adopters":0,"source":"github","repoUrl":"https://github.com/steipete/agent-rules","stars":5684,"score":88,"safety":72,"grade":"S","factors":[{"name":"Documentation","points":15,"max":20,"detail":"3/3 required SKILL.md sections, 320 chars"},{"name":"Capability breadth","points":15,"max":15,"detail":"4 declared capabilities"},{"name":"Agent compatibility","points":15,"max":15,"detail":"Works with openclaw, claude-code, cursor, generic-mcp"},{"name":"License openness","points":10,"max":10,"detail":"MIT — reuse freely"},{"name":"Community signal","points":20,"max":20,"detail":"5684 GitHub stars"},{"name":"Safety posture","points":13,"max":20,"detail":"Risk surface: network-egress, external-api"}],"seal":"06ac2e68fca85f3311f02edb50f98f0a9a15e3d4c146277cc6f02e68c90adc65d574fa32c063c9456b3bb6369e89877a","sealedAt":"2026-09-21T18:41:18.860Z"},{"id":"gh-1002608585","name":"mcp-documentation-server","tagline":"MCP Documentation Server - Bridge the AI Knowledge Gap. ✨ Features: Document man","description":"MCP Documentation Server - Bridge the AI Knowledge Gap. ✨ Features: Document management • Gemini integration • AI-powered semantic search • File uploads • Smart","category":"research","capabilities":["documents","gemini","knowledge base","mcp server"],"worksWith":["openclaw","claude-code","cursor","generic-mcp"],"skillMd":"# mcp-documentation-server\n\nCommunity index (GitHub): https://github.com/andrea9293/mcp-documentation-server\n\n## Triggers\nLooking for reusable research workflows.\n\n## Workflow\n1. Read the project README.\n2. Extract the reusable pattern.\n3. Adapt it into a local SKILL.md.\n\n## Guardrails\n- Audit code before running.\n- Pin versions; review permissions.\n","version":"0.live","author":"andrea9293","license":"MIT","riskFlags":["network-egress","external-api"],"adopters":0,"source":"github","repoUrl":"https://github.com/andrea9293/mcp-documentation-server","stars":344,"score":88,"safety":72,"grade":"S","factors":[{"name":"Documentation","points":15,"max":20,"detail":"3/3 required SKILL.md sections, 352 chars"},{"name":"Capability breadth","points":15,"max":15,"detail":"4 declared capabilities"},{"name":"Agent compatibility","points":15,"max":15,"detail":"Works with openclaw, claude-code, cursor, generic-mcp"},{"name":"License openness","points":10,"max":10,"detail":"MIT — reuse freely"},{"name":"Community signal","points":20,"max":20,"detail":"344 GitHub stars"},{"name":"Safety posture","points":13,"max":20,"detail":"Risk surface: network-egress, external-api"}],"seal":"cee86ca98c820b1636fb03d4de1cf461668bae41b915661c21b04d46923b66310bc8a1e07df75d1d1b07ee379d80e39f","sealedAt":"2026-09-21T18:41:18.860Z"}],"total":10024,"page":1,"pages":418,"perPage":24,"catalog":10024,"live":10000,"head":"e800631e36483c866a9a2ba7ae87b272f194cb0f2289e715f925a86e78361595ad238e6f39fa17c874239ccb7bd04a8d","facets":{"byCategory":{"research":903,"code":4987,"data":401,"docs":264,"media":387,"ops":3082},"bySource":{"curated":24,"github":2809,"mcp":7,"npm":7184}},"generatedAt":"2026-09-21T18:41:18.860Z","algo":"SHA-384(prevSeal || canonicalJson)"}