universal-skills-marketplace

Universal Skills Marketplace

Context7 for skills, not docs. One MCP server. Two ecosystems. Unofficial.

Mission

Build, maintain, or extend an unofficial bridge that lets any Claude Code or OpenAI Codex session search one registry for skills originating in either ecosystem, get them back with quality scores + install commands for both CLIs, and load them on demand.

Not affiliated with Anthropic or OpenAI. Identity details in references/00-architecture-overview.md.

Routing table — pick the right reference

I want to…	Primary reference	Supporting asset / script
Understand the whole system	references/00-architecture-overview.md	assets/diagrams/system-topology.png
Learn the SKILL.md open standard	references/01-agentskills-io-spec-walkthrough.md	assets/templates/standalone-SKILL.md.template
Author a Claude plugin	references/02-claude-plugin-format.md	assets/templates/claude-plugin.json.template, assets/real-examples/context7-plugin.json
Author a Codex plugin	references/03-codex-plugin-format.md	assets/templates/codex-plugin.json.template, assets/real-examples/openai-canva-plugin.json
Design MCP tools	references/04-mcp-tool-design.md	assets/templates/mcp-server-index.ts.template
Deploy Cloudflare Workers	references/05-cloudflare-workers-playbook.md	assets/templates/wrangler.toml.*.template
Design the D1 catalog	references/06-d1-schema-design.md	assets/templates/d1-schema.sql.template
Plan R2 content storage	references/07-r2-storage-patterns.md	—
Build the upstream indexer	references/08-github-indexer-design.md	scripts/fetch-upstream-catalog.sh
Score skill quality 0-100	references/09-quality-scoring-rubric.md	assets/fixtures/known-good/
Copy what Context7 got right	references/10-context7-architectural-analysis.md	assets/real-examples/context7-*.json
Translate between manifest formats	references/11-manifest-translator-algorithm.md	scripts/test-translator.ts, assets/fixtures/lossy-cases/
Verify end-to-end	references/12-verification-playbook.md	scripts/validate.sh

Phase dispatcher

• Planning / just arrived → read 00-architecture-overview.md, then 10-context7-architectural-analysis.md, then references/11-manifest-translator-algorithm.md (the three anchoring docs)
• Phase 0 — spikes complete → evidence in /docs/spikes/ (upstream URLs verified, Codex schema derived, rate limits mapped, iCloud strategy documented)
• Phase 1 — schemas + this skill → you are here; running this skill IS Phase 1's deliverable
• Phase 2 — npm package → packages/mcp-server/ implementation driven by 04-mcp-tool-design.md + 11-manifest-translator-algorithm.md
• Phase 3 — Cloudflare backend → three Workers driven by 05-cloudflare-workers-playbook.md + 06-d1-schema-design.md + 08-github-indexer-design.md

Hard rules (inherited from ../../CLAUDE.md)

1. Anti-Fabrication — if a URL, field, or behavior isn't verified against installed plugins or a live upstream, it does NOT ship.
2. Anti-Mock-Data — no hardcoded skills anywhere except assets/fixtures/ and packages/*/tests/fixtures/.
3. Verification — "build passed" ≠ acceptance; curl + fresh Claude Code AND Codex session required.
4. Translator loudness — no silent field drops; every lossy translation logs + shims.
5. context7 fidelity — plugin wrapper stays trivial; all logic in the npm package.
6. Identity — unofficial, independent, community project. Never imply endorsement from either company.

Run before any commit

bash skills/universal-skills-marketplace/scripts/validate.sh

Must exit 0. Checks: frontmatter valid, SKILL.md ≤100 lines, all 12 references present, all templates parse, no broken intra-skill links.