Claude Code Plugins Plus Skills

{
  "service_type": "platform",
  "base_url": "https://tonsofskills.com",
  "auth_method": "none",
  "auth_config": {},
  "endpoints": [],
  "pricing_model": {
    "type": "free",
    "details": {
      "model": "open-source marketplace",
      "notes": "Free to browse and use; plugins and skills are open-source. No paid tiers indicated."
    }
  },
  "rate_limits": {},
  "capabilities": [
    "Plugin marketplace for Claude Code (340 plugins)",
    "Agent skills library (1367 agent skills)",
    "CCPI package manager for installing plugins/skills",
    "Interactive tutorials for onboarding and usage",
    "Production orchestration patterns for agents",
    "Web-based discovery and browsing interface"
  ],
  "raw_analysis": "Claude Code Plugins Plus Skills (tonsofskills.com) is a community/open-source marketplace and resource hub for the Claude Code ecosystem. It aggregates 340 plugins and 1367 agent skills, providing a discovery and distribution layer for reusable Claude Code extensions. The site includes the CCPI package manager, which appears to be a CLI or tool for installing and managing these packages within a Claude Code environment, along with interactive tutorials and documented production orchestration patterns. \n\nTarget audience: developers and teams building with Claude Code / AI agent workflows — from individuals looking for productivity skills to teams implementing production-grade agent orchestration. \n\nMaturity: Appears to be a relatively new niche community project rather than a large commercial platform. The listing as 'api' category is slightly misleading, as no public REST API is documented on the surface; the primary interaction model is likely a website plus CLI package manager. No authentication, pricing tiers, or rate limits are evident from available information. \n\nIntegrations: Primarily integrates with Claude Code (Anthropic's CLI/agent tooling) and its plugin/skill system. Distribution is via the CCPI package manager. It is not itself an API-first service; its value is as a curated registry and knowledge base. Because no public REST endpoints, auth model, or formal API documentation were found in the provided content, endpoints and rate limits are left empty pending direct site inspection or documentation review."
}

2/3 tests passed

```json
{
  "overall": 71,
  "dimensions": {
    "token_efficiency": 9.0,
    "first_try_success": 8.0,
    "response_parseability": 6.0,
    "error_clarity": 6.0,
    "doc_quality": 7.5,
    "auth_simplicity": 10.0,
    "latency": 9.5,
    "consistency": 7.5
  },
  "pricing_normalized": {
    "model": "free",
    "tier": "open-source marketplace",
    "cost_per_month_usd": 0,
    "notes": "Free to browse and use; plugins and skills are open-source. No paid tiers indicated."
  },
  "issues": [
    "No /llms.txt endpoint (404) — agents lack a machine-readable self-description of the platform",
    "No API surface detected — discovery is HTML-browsing only, which is token-heavy for agents",
    "Sitemap exists but no structured JSON feed of plugins/skills for programmatic consumption",
    "Naming counts are inconsistent across surfaces (340 plugins/1367 skills vs '2,900 marketplace' in meta description), which can confuse agent summaries",
    "Error pages are styled HTML 404s with no structured error guidance for non-human consumers"
  ],
  "recommendations": [
    "Publish /llms.txt and /llms-full.txt describing the marketplace, CCPI package manager, and how agents should discover/install plugins",
    "Expose a JSON API or static JSON index (e.g., /api/plugins.json, /api/skills.json) so agents can query catalog metadata without scraping HTML",
    "Add an OpenAPI or schema.org JSON-LD description per plugin/skill for structured discovery",
    "Reconcile catalog counts (plugins vs skills vs total) into a single canonical source referenced by meta tags and docs",
    "Provide a machine-parseable error contract for 404s (e.g., JSON when Accept: application/json)",
    "Document the CCPI install flow with copy-pasteable commands and exit/error semantics",
    "Add a status/health endpoint and uptime badge so agents can verify reliability before recommending"
  ]
}
```