CCM
/MCP
SkillsMCPMarketplacesDigestToolsAdvertise

This week in Claude

Every Monday: Claude Code, Agent SDK, MCP, and the Anthropic platform moves worth your time.

Skills by Category
Frontend DevelopmentBackend & APIsTesting & QASecurityDevOps & CI/CDGit & Pull RequestsDocumentationCode Review & QualityAI & Agent BuildingSkill Development
MCP Servers by Category
Sales & MarketingWeb & Browser AutomationDatabasesAI & LLM ToolsCloud & InfrastructureCommunication & MessagingDeveloper ToolsDesign & CreativeDocuments & KnowledgeSearch & Web Crawling
Marketplaces by Category
AI Agents & OrchestrationLLM IntegrationDevelopment ToolsFrontend & UIBackend & APIsDatabasesTesting & Code QualityDevOps & CloudSecurity & ComplianceGit & Version Control

Claude Code Marketplaces

Discover Claude Code plugins, extensions, and tools. Automatically updated directory of Anthropic Claude AI marketplaces with development tools, productivity plugins, and integrations.

Resources

  • Browse Skills
  • Browse MCP Servers
  • Browse Marketplaces
  • Skill index
  • MCP index
  • Marketplace index
  • Plugins Reference

Community

  • About
  • Tools
  • Feedback
  • Privacy Policy
  • Advertise

Built for the Claude Code community with Claude Code by mertbuilds.com

Independent project, not affiliated with Anthropic
digicatalyst-systems avatar

Dep Diff Mcp

digicatalyst-systems/dep-diff-mcp
22 toolsSTDIOregistry active
Summary

Point this at a Dependabot PR or npm outdated output and get a ranked upgrade plan with breaking changes, CVE fixes, and migration links. The server exposes two tools: analyze_package_change for single packages and analyze_packages_bulk for up to 50 at once. It scrapes GitHub release notes, checks OSV.dev for security fixes, and returns a risk level (security, caution, review, likely-safe, safe) with semver classification. Handles npm and PyPI. Runs via npx with optional GitHub token for higher rate limits. Integrates with Claude Code, Cursor, and Claude Desktop. Good for turning lockfile diffs into actionable decisions without manually reading changelogs.

CodeRabbit
CodeRabbit
AI writes the code. CodeRabbit catches the slop.
Try For Free →
ego lite browserego lite browser
ego lite browser
Fastest browser for AI agents to run web automation tasks, always free.
Download Free life-time →
Granola, the best AI meeting recorder
Granola, the best AI meeting recorder
Notes, actions and memory. Without a meeting bot. First month 100% off.
Download for free →
CodeHealth MCP ServerCodeHealth MCP Server
CodeHealth MCP Server
Protect your code quality, stop the AI slop.
Try For Free →
belt - the only tool your agent needs
belt - the only tool your agent needs
belt cli automatically finds the best tools and skills for your agent. image, video, music, tts...
one prompt install →
AppSignal
AppSignal
Monitor with ease. Code with confidence.
Start Free Trial →
Agent, connect blockchain
Agent, connect blockchain
Connect your Claude agent to live crypto prices and trading routes via 1inch
Get the MCP →
Block distraction from your iPhone for freeBlock distraction from your iPhone for free
Block distraction from your iPhone for free
Block distracting apps from your iPhone permanently without a 3rd party app. Free and open source.
Block now (100% free) →
CodeRabbit
CodeRabbit
AI writes the code. CodeRabbit catches the slop.
Try For Free →
ego lite browserego lite browser
ego lite browser
Fastest browser for AI agents to run web automation tasks, always free.
Download Free life-time →
Granola, the best AI meeting recorder
Granola, the best AI meeting recorder
Notes, actions and memory. Without a meeting bot. First month 100% off.
Download for free →
CodeHealth MCP ServerCodeHealth MCP Server
CodeHealth MCP Server
Protect your code quality, stop the AI slop.
Try For Free →
belt - the only tool your agent needs
belt - the only tool your agent needs
belt cli automatically finds the best tools and skills for your agent. image, video, music, tts...
one prompt install →
AppSignal
AppSignal
Monitor with ease. Code with confidence.
Start Free Trial →
Agent, connect blockchain
Agent, connect blockchain
Connect your Claude agent to live crypto prices and trading routes via 1inch
Get the MCP →
Block distraction from your iPhone for freeBlock distraction from your iPhone for free
Block distraction from your iPhone for free
Block distracting apps from your iPhone permanently without a 3rd party app. Free and open source.
Block now (100% free) →

Tools

Public tool metadata for what this MCP can expose to an agent.

2 tools
analyze_package_changeGiven one package and two versions (from -> to), returns a structured upgrade analysis: semver classification, GitHub release notes summary, detected breaking changes, security advisories fixed in the range, migration guide links, and a clear recommendation. Use when the user...4 params

Given one package and two versions (from -> to), returns a structured upgrade analysis: semver classification, GitHub release notes summary, detected breaking changes, security advisories fixed in the range, migration guide links, and a clear recommendation. Use when the user...

Parameters* required
ecosystemstring
Package ecosystemone of npm · pypi
fromVersionstring
Current version (e.g. '18.2.0')
namestring
Package name (e.g. 'react', 'requests')
toVersionstring
Target version (e.g. '19.0.0')
analyze_packages_bulkAnalyzes a list of package upgrades in parallel and returns a unified risk report with packages ranked by recommendation level (security > caution > review > likely-safe > safe). Use when the user provides many dependency changes from a Dependabot PR, npm outdated output, lock...1 params

Analyzes a list of package upgrades in parallel and returns a unified risk report with packages ranked by recommendation level (security > caution > review > likely-safe > safe). Use when the user provides many dependency changes from a Dependabot PR, npm outdated output, lock...

Parameters* required
changesarray
List of package changes to analyze

dep-diff-mcp

MCP server that translates a lockfile diff into a human-readable upgrade plan.

Point your AI assistant (Cursor, Claude Desktop, Claude Code) at a Dependabot PR, npm outdated output, or any pair of package versions, and get back a ranked upgrade plan: semver class, breaking changes pulled from GitHub release notes, CVEs fixed in the range, migration guide links, and a clear recommendation per package.

What it looks like

Ask your assistant, in plain language:

Is it safe to bump lodash from 4.17.20 to 4.17.21?

It calls analyze_package_change and gets back:

{
  "package": "lodash",
  "ecosystem": "npm",
  "fromVersion": "4.17.20",
  "toVersion": "4.17.21",
  "semverClass": "patch",
  "repoUrl": "https://github.com/lodash/lodash",
  "releaseCount": 0,
  "breakingChanges": [],
  "securityFixes": [
    {
      "id": "GHSA-29mw-wpgm-hmr9",
      "summary": "Regular Expression Denial of Service (ReDoS) in lodash",
      "severity": "MODERATE"
    },
    {
      "id": "GHSA-35jh-r3h4-6jhm",
      "summary": "Command Injection in lodash",
      "severity": "HIGH"
    }
  ],
  "migrationLinks": [],
  "recommendation": "RECOMMENDED: 2 security fix(es) (incl. high/critical).",
  "recommendationLevel": "security"
}

A patch bump you would normally merge without looking. It closes a HIGH-severity command injection. That is the case this server exists for.

GitHub Actions

Dependabot opens more actions/* pull requests than almost anything else, and its scheduled version updates carry no advisory data. Ask:

Is it safe to bump tj-actions/changed-files from 45.0.7 to 46.0.1?

{
  "package": "tj-actions/changed-files",
  "ecosystem": "github-actions",
  "semverClass": "major",
  "repoUrl": "https://github.com/tj-actions/changed-files",
  "securityFixes": [
    {
      "id": "GHSA-mrrh-fwg8-r2c3",
      "summary": "tj-actions changed-files through 45.0.7 allows remote attackers to discover secrets",
      "severity": "HIGH"
    }
  ],
  "recommendation": "RECOMMENDED: 1 security fix(es) (incl. high/critical).",
  "recommendationLevel": "security"
}

A whole Dependabot batch

Here's my Dependabot PR — what's actually risky in it?

analyze_packages_bulk takes up to 50 changes at once and ranks them security > caution > review > likely-safe > safe:

{
  "totalPackages": 5,
  "bySemverClass": { "major": 1, "minor": 3, "patch": 1 },
  "securityFixesTotal": 7,
  "packagesWithBreakingChanges": 1,
  "packages": [ /* one entry per package, same shape as above */ ]
}

Condensing the recommendation field of each entry, that batch comes back in this order:

PackageChangeClassVerdict
lodash4.17.20 → 4.17.21patchRECOMMENDED — 2 security fixes (incl. high/critical)
axios1.6.0 → 1.7.9minorRECOMMENDED — 1 security fix (incl. high/critical)
express4.18.2 → 5.0.0majorRECOMMENDED — 2 security fixes; 3 breaking changes, migration guide
requests (PyPI)2.31.0 → 2.32.0minorRECOMMENDED — 2 security fixes
typescript5.3.3 → 5.4.5minorLIKELY SAFE — minor version, additive changes per semver

Note the ordering: the patch bump outranks the major one. Semver tells you how much changed; it does not tell you what is urgent.

Every response above is real output from the hosted instance, trimmed only where marked.

Install

Claude Code

One command, user scope (available in every project):

claude mcp add -s user dep-diff -- npx -y @digicatalyst/dep-diff-mcp

Project scope (writes .mcp.json at repo root, team-shared):

claude mcp add -s project dep-diff -- npx -y @digicatalyst/dep-diff-mcp

With an explicit token (skip this if you have the gh CLI authenticated — see GitHub token below):

claude mcp add -s user --env GITHUB_TOKEN=ghp_xxx dep-diff -- npx -y @digicatalyst/dep-diff-mcp

Verify:

claude mcp list

Restart the Claude Code session to pick up the server.

Cursor and Claude Desktop

Add to your MCP client config:

  • Cursor: ~/.cursor/mcp.json
  • Claude Desktop: ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows)
{
  "mcpServers": {
    "dep-diff": {
      "command": "npx",
      "args": ["-y", "@digicatalyst/dep-diff-mcp"]
    }
  }
}

Restart your MCP client. Ask something like "what's risky in this Dependabot PR?" and the tools are invoked automatically.

Hosted remote (no install)

A hosted instance runs at https://dep-diff.digicatalyst.ca/mcp over streamable HTTP, so you can skip the npm package entirely:

claude mcp add -s user -t http dep-diff https://dep-diff.digicatalyst.ca/mcp

Or in a client config:

{
  "mcpServers": {
    "dep-diff": {
      "type": "http",
      "url": "https://dep-diff.digicatalyst.ca/mcp"
    }
  }
}

The hosted instance authenticates to GitHub with its own token, so release-note lookups work at full rate limits without you configuring anything. Pass ?githubToken=ghp_xxx on the URL only if you would rather requests counted against your own GitHub quota. It is stateless and keeps no logs of your queries — see PRIVACY.md. Run the npm package locally instead if you would rather your token never leave your machine.

The same instance is also listed on Smithery, which proxies to it through their gateway.

GitHub token (optional but recommended)

The server hits the GitHub API to read release notes. Without a token you get 60 requests per hour (GitHub's anonymous limit) — enough for occasional single-package queries, not enough for bulk lockfile analysis.

The server resolves a token in this order:

  1. GITHUB_TOKEN environment variable, if set.
  2. gh auth token — if the GitHub CLI is installed and authenticated, the server uses that token automatically. No config change needed.
  3. Anonymous (60 req/hr).

Recommended: use the gh CLI

If you already have gh installed (brew install gh && gh auth login), stop here — the server picks up your existing auth. No plaintext token anywhere.

Alternative: environment variable

Create a fine-grained token at https://github.com/settings/tokens:

  • Token name: dep-diff-mcp
  • Expiration: 90 days (rotate periodically)
  • Repository access: Public Repositories (read-only) — no private repo access
  • Permissions: none beyond the default public read — do not grant repo, workflow, user, or any write scope

Then reference it in the MCP config:

{
  "mcpServers": {
    "dep-diff": {
      "command": "npx",
      "args": ["-y", "@digicatalyst/dep-diff-mcp"],
      "env": { "GITHUB_TOKEN": "github_pat_xxx" }
    }
  }
}

Security notes

  • This config file lives on your disk in plaintext. Keep perms tight (chmod 600) and do not paste the token into AI chats, issues, or shared screens — transcripts are often retained.
  • The token in this config should be least-privilege (public repo read only). Even leaked, it can only read public data you could already read.
  • Rotate tokens periodically. Revoke any token that may have been exposed at https://github.com/settings/tokens.
  • The server never writes the token to stdout/stderr or the response payload.

Tools

analyze_package_change

Analyze one package upgrade. Inputs: ecosystem (npm, pypi, or github-actions), name, fromVersion, toVersion.

analyze_packages_bulk

Analyze up to 50 package upgrades in parallel. Returns packages ranked by risk (security > caution > review > likely-safe > safe), plus summary counts.

What you get back

  • Semver classification — major / minor / patch / downgrade / unknown
  • Breaking changes — extracted from GitHub release notes headers
  • Security fixes — CVEs present at fromVersion but resolved at toVersion (via OSV.dev)
  • Migration links — upgrade guide URLs found in release notes
  • Recommendation — single-line verdict + level

Both tools return this twice: as the JSON text block shown above, and as MCP structured content validated against a declared outputSchema. A client that supports structured output can read recommendationLevel or securityFixes[] straight off the response instead of re-parsing the text. Clients that don't are unaffected — the text block is unchanged.

Supported ecosystems

  • npm
  • PyPI
  • GitHub Actions — pass the action reference as the name (actions/checkout, or a nested github/codeql-action/init). An action reference is already a repository coordinate, so release notes and breaking changes resolve without a registry lookup, and bare major tags such as 4 -> 5 classify correctly.

Development

npm install
npm run build
GITHUB_TOKEN=ghp_xxx npm run inspect   # MCP Inspector

License

MIT

Featured
CodeRabbit
CodeRabbit
AI writes the code. CodeRabbit catches the slop.
Try For Free →
ego lite browserego lite browser
ego lite browser
Fastest browser for AI agents to run web automation tasks, always free.
Download Free life-time →
Granola, the best AI meeting recorder
Granola, the best AI meeting recorder
Notes, actions and memory. Without a meeting bot. First month 100% off.
Download for free →
CodeHealth MCP ServerCodeHealth MCP Server
CodeHealth MCP Server
Protect your code quality, stop the AI slop.
Try For Free →
belt - the only tool your agent needs
belt - the only tool your agent needs
belt cli automatically finds the best tools and skills for your agent. image, video, music, tts...
one prompt install →
AppSignal
AppSignal
Monitor with ease. Code with confidence.
Start Free Trial →
Agent, connect blockchain
Agent, connect blockchain
Connect your Claude agent to live crypto prices and trading routes via 1inch
Get the MCP →
Block distraction from your iPhone for freeBlock distraction from your iPhone for free
Block distraction from your iPhone for free
Block distracting apps from your iPhone permanently without a 3rd party app. Free and open source.
Block now (100% free) →
Registryactive
Package@digicatalyst/dep-diff-mcp
TransportSTDIO
UpdatedApr 23, 2026
View on GitHub