Skip to content

Releases: adityasugandhi/skillsync-mcp

v1.4.1

Choose a tag to compare

@adityasugandhi adityasugandhi released this 23 Feb 02:53

Improve Socket.dev Supply Chain Security Score

  • Remove @scarf/scarf dependency and scarfSettings (eliminates obfuscated code + telemetry HIGH alerts)
  • Escape eval regex pattern in patterns.ts to avoid "uses eval" false positive
  • Exclude build/__tests__/ from published npm tarball
  • Bump version to 1.4.1

v1.4.0 — Multi-scope skill management

Choose a tag to compare

@adityasugandhi adityasugandhi released this 23 Feb 02:22

What's New

  • Multi-scope skill management: Support for both global (~/.claude/) and project-scoped (./.claude/) skills
  • Scope resolver: New scope-resolver.ts for path resolution and validation across scopes
  • All tools updated: Every MCP tool now accepts an optional scope parameter (defaults to global for backward compatibility)
  • CI/CD pipeline: GitHub Actions workflow for build/test (Node 20/22) and automated npm publish on release
  • 135 tests passing across security scanner, skill manager, sync engine, and scope resolver

Full Changelog

v1.3.1...v1.4.0

v1.3.0 — Community Launch

Choose a tag to compare

@adityasugandhi adityasugandhi released this 21 Feb 20:03

What's New

8 MCP Tools

  • skillsmp_search — Keyword search across SkillsMP marketplace
  • skillsmp_ai_search — AI-powered semantic search (Cloudflare AI)
  • skillsmp_scan_skill — Security scan a GitHub skill repo (60+ patterns)
  • skillsmp_search_safe — Search + auto-scan top results
  • skillsmp_install_skill — Scan, gate, and install to ~/.claude/skills/
  • skillsmp_uninstall_skill — Remove an installed skill
  • NEW skillsmp_list_installed — List installed skills with risk levels
  • NEW skillsmp_audit_installed — Deep security audit of installed skills

Security Scanner

  • 95 threat patterns across 20 categories
  • Critical patterns block installation unconditionally
  • Multi-line evasion detection
  • Content hashing for TOCTOU verification
  • Output sanitization against prompt injection

Cloud Sync Engine

  • Subscription-based skill sync from SkillsMP API
  • Lock file for stable change detection
  • Conflict policies (skip/overwrite)
  • Dry run support

Multi-Client Support

  • Claude Code, OpenClaw, Cursor, Windsurf, nanobot
  • Configuration examples for all clients

Community Infrastructure

  • Issue templates (bug report, feature request, skill submission)
  • PR template with checklist
  • CONTRIBUTING.md with contribution guides
  • SKILL.md for agent discoverability
  • docs/THREAT_PATTERNS.md documenting all 95 patterns

Install

{
  "mcpServers": {
    "skillsmp": {
      "command": "npx",
      "args": ["-y", "@stranzwersweb2/skillsync-mcp"]
    }
  }
}