Atlas / MCP servers / musaceylan / Claude Intel

Claude IntelSAFE

Name: Claude Intel
Author: musaceylan

mcp/musaceylan/claude-intel

MCP server by musaceylan · Audited 2026-10-09

AI-powered MCP server that continuously improves Claude Code by learning from high-signal GitHub repositories — auto-evolving developer intelligence layer

Verdict

SAFE

Grade

Trust score

89 /100

Exposed tools

18 15r · 3w · 0d

Transport

stdio

License

—

Stars

Overview

From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.

An auto-evolving developer intelligence layer for Claude Code.

claude-intel-mcp continuously scans high-signal GitHub repositories, extracts patterns, compares them against your local project, and proposes safe improvements to your CLAUDE.md — keeping your Claude Code setup ahead of the curve.

Architecture

┌──────────────────────────────────────────────────────────────┐
│                     claude-intel-mcp                         │
│                                                              │
│  ┌─────────────────────────────────────────────────────┐    │
│  │  MCP Server (stdio)                                  │    │
│  │   ├── Tools (7)    ├── Resources (7)  ├── Prompts (2)│    │
│  └──────────────────────────────────────────────────────┘    │
│            │                                                 │
│  ┌─────────▼─────────────────────────────────────┐          │
│  │  Core Engine                                   │          │
│  │   ├── GithubClient (API + rate limiting)       │          │
│  │   ├── GithubCollector (multi-query search)     │          │
│  │   ├── RepoClassifier (hype/abandoned/mcp...)   │          │
│  │   ├── RelevanceScorer (5-factor, 0-1 score)    │          │
│  │   ├── PatternExtractor (CLAUDE.md/MCP/config)  │          │
│  │   ├── PatternMemory (SQLite, deduped)          │          │
│  │   ├── DnaProfiler (local stack detection)      │          │
│  │   ├── Comparator (gap/enhancement analysis)   │          │
│  │   ├── SuggestionEngine (micro/meso/macro)      │          │
│  │   ├── PatchBuilder (safe unified diffs)        │          │
│  │   ├── UpdatePlanner (phased update plans)      │          │
│  │   └── LearningLog (SQLite audit trail)         │          │
│  └───────────────────────────────────────────────┘          │
│                                                              │
│  ┌─────────────────────────────┐                            │
│  │  Storage (~/.claude-intel/) │

Read from source at commit 83b3c22be5e3OBSERVED · 2026-10-09

Connect

Built from this server's own package name, version and transport as found in its source — not copied from anyone's documentation, so it cannot drift against a page we do not control.

claude-code

claude mcp add claude-intel-mcp -- npx -y [email protected]

claude-desktop

{
  "mcpServers": {
    "claude-intel-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "[email protected]"
      ]
    }
  }
}

Exposed tools (18)

15 read · 3 write · 0 destructive.

Tool	Risk	Description
`apply_safe_updates`	write	Apply a safe, non-destructive update to CLAUDE.md. Requires explicit confirm=true. Creates a backup before applying.
`compare_files`	read	Compare files
`compare_with_local_repo`	read	Compare an external GitHub repository
`create_update_plan`	write	Create a structured update plan from a list of suggestions, grouped into phases.
`focus`	read	Specific sections to focus on
`gaps`	read	Identified gaps
`generate_claude_md_patch`	write	Generate a unified diff patch for CLAUDE.md based on improvement suggestions. Use dryRun=true (default) to preview without writing.
`generate_report`	read
`generate_suggestions`	read	Generate improvement suggestions by running a full scan → classify → compare pipeline against the local repo.
`mcp-server-template`	read	A boilerplate template for MCP servers
`rank_repositories`	read	Rank a list of GitHub repositories by their relevance to Claude Code development. Returns repos sorted by score with detailed breakdown.
`repos`	read	Ranked repos to summarize
`rewrite_claude_md`	read	Generate a prompt to ask Claude to rewrite/improve CLAUDE.md incorporating suggestions from the analysis
`scan_github_trends`	read	Scan GitHub for high-signal repositories related to Claude Code, MCP servers, and AI developer workflows. Returns ranked repos with relevance scores.
`scan_repos`	read	Scan repositories
`suggestions`	read	Brief list of suggested improvements
`summarize_insights`	read	Generate a prompt to summarize all findings from the scan and analysis
`topRepos`	read	Names of top-scoring repos to use as reference

Trust audit

SAFEgrade B · trust 89/100 Nothing in the source contradicts what it says it does. Grade A is reserved for packages that have also passed the behavioural sandbox.

Layer	What it checks	Result
L0	Provenance & inventory	PASS
L1	Static analysis of the code	PASS
L2	Instruction surface (what it tells the agent)	PASS
L3	Class-specific surface	PASS
L4	Behavioural (sandbox)	SKIPPED

What the source does

Filesystem: UNDECLARED (5 observation(s))
Network: declared (1 observation(s))
Shell: none-observed
Dependencies: not all pinned
Secrets in source: none-found