Claude-skill-registry dev-dependency-management

Package and dependency management patterns across ecosystems (npm, pip, cargo, maven). Covers lockfiles, semantic versioning, dependency security scanning, update strategies, monorepo workspaces, transitive dependencies, and avoiding dependency hell.

install
source · Clone the upstream repo
git clone https://github.com/majiayu000/claude-skill-registry
Claude Code · Install into ~/.claude/skills/
T=$(mktemp -d) && git clone --depth=1 https://github.com/majiayu000/claude-skill-registry "$T" && mkdir -p ~/.claude/skills && cp -r "$T/skills/data/dev-dependency-management" ~/.claude/skills/majiayu000-claude-skill-registry-dev-dependency-management && rm -rf "$T"
manifest: skills/data/dev-dependency-management/SKILL.md
source content

Dependency Management — Production Patterns

Modern Best Practices (January 2026): Lockfile-first workflows, automated security scanning (Dependabot, Snyk, Socket.dev), semantic versioning, minimal dependencies principle, monorepo workspaces (pnpm, Nx, Turborepo), supply chain security (SBOM, AI BOM, Sigstore), reproducible builds, and AI-generated code validation.


When to Use This Skill

The agent should invoke this skill when a user requests:

  • Adding new dependencies to a project
  • Updating existing dependencies safely
  • Resolving dependency conflicts or version mismatches
  • Auditing dependencies for security vulnerabilities
  • Understanding lockfile management and reproducible builds
  • Setting up monorepo workspaces (pnpm, npm, yarn)
  • Managing transitive dependencies and overrides
  • Choosing between similar packages (bundle size, maintenance, security)
  • Dependency version constraints and semantic versioning
  • Dependency security best practices and supply chain security
  • Troubleshooting "dependency hell" scenarios
  • Package manager configuration and optimization
  • Creating reproducible builds across environments

Quick Reference

TaskTool/CommandKey ActionWhen to Use
Install from lockfile
npm ci
,
poetry install
,
cargo build
Clean install, reproducibleCI/CD, production deployments
Add dependency
npm install <pkg>
,
poetry add <pkg>
Updates lockfile automaticallyNew feature needs library
Update dependencies
npm update
,
poetry update
,
cargo update
Updates within version constraintsMonthly/quarterly maintenance
Check for vulnerabilities
npm audit
,
pip-audit
,
cargo audit
Scans for known CVEsBefore releases, weekly
View dependency tree
npm ls
,
pnpm why
,
pipdeptree
Shows transitive dependenciesDebugging conflicts
Override transitive dep
overrides
(npm),
pnpm.overrides
Force specific versionSecurity patch, conflict resolution
Monorepo setup
pnpm workspaces
,
npm workspaces
Shared dependencies, cross-linkingMulti-package projects
Check outdated
npm outdated
,
poetry show --outdated
Lists available updatesPlanning update sprints

Decision Tree: Dependency Management

User needs: [Dependency Task]
    ├─ Adding new dependency?
    │   ├─ Check: Do I really need this? (Can implement in <100 LOC?)
    │   ├─ Check: Is it well-maintained? (Last commit <6 months, >10k downloads/week)
    │   ├─ Check: Bundle size impact? (Use Bundlephobia for JS)
    │   ├─ Check: Security risks? (`npm audit`, Snyk)
    │   └─ If all checks pass → Add with `npm install <pkg>` → Commit lockfile
    │
    ├─ Updating dependencies?
    │   ├─ Security vulnerability? → `npm audit fix` → Test → Deploy immediately
    │   ├─ Routine update?
    │       ├─ Patch versions → `npm update` → Safe, do frequently
    │       ├─ Minor/major → Check CHANGELOG → Test in staging → Update gradually
    │       └─ All at once → [FAIL] RISKY → Update in batches instead
    │
    ├─ Dependency conflict?
    │   ├─ Transitive dependency issue?
    │       ├─ View tree: `npm ls <package>`
    │       ├─ Use overrides sparingly: `overrides` in package.json
    │       └─ Document why override is needed
    │   └─ Peer dependency mismatch?
    │       └─ Check version compatibility → Update parent or child
    │
	├─ Monorepo project?
	│   ├─ Use pnpm workspaces (recommended default)
	│   ├─ Shared deps → Root package.json
	│   ├─ Package-specific → Package directories
	│   └─ Use Nx or Turborepo for task caching
	│
	└─ Choosing package manager?
	    ├─ New JS project → **pnpm** (recommended default) or **Bun** (often faster; verify ecosystem maturity)
	    ├─ Enterprise monorepo → **pnpm** (mature workspace support)
	    ├─ Speed-focused experimentation → **Bun** (verify ecosystem maturity)
	    ├─ Existing npm project → Migrate to pnpm or stay (check team preference)
	    ├─ Python → **uv** (fast), Poetry (mature), pip+venv (simple)
	    └─ Data science → **conda** or **uv** (faster environment setup)

Navigation: Core Patterns

Lockfile Management

references/lockfile-management.md

Lockfiles ensure reproducible builds by recording exact versions of all dependencies (direct + transitive). Essential for preventing "works on my machine" issues.

  • Golden rules (always commit, never edit manually, regenerate on changes)
  • Commands by ecosystem (npm ci, poetry install, cargo build)
  • Troubleshooting lockfile conflicts
  • CI/CD integration patterns

Semantic Versioning (SemVer)

references/semver-guide.md

Understanding version constraints (

^
,
~
, exact) and how to specify dependency ranges safely.

  • SemVer format (MAJOR.MINOR.PATCH)
  • Version constraint syntax (caret, tilde, exact)
  • Recommended strategies by project type
  • Cross-ecosystem version management

Dependency Security Auditing

references/security-scanning.md

Automated security scanning, vulnerability management, and supply chain security best practices.

  • Automated tools (Dependabot, Snyk, GitHub Advanced Security)
  • Running audits (npm audit, pip-audit, cargo audit)
  • CI integration and alert configuration
  • Incident response workflows

Dependency Selection

references/dependency-selection-guide.md

Deciding whether to add a new dependency and choosing between similar packages.

  • Minimal dependencies principle (best dependency is the one you don't add)
  • Evaluation checklist (maintenance, bundle size, security, alternatives)
  • Choosing between similar packages (comparison matrix)
  • When to reject a dependency

Update Strategies

references/update-strategies.md

Keeping dependencies up to date safely while minimizing breaking changes and security risks.

  • Update strategies (continuous, scheduled, security-only)
  • Safe update workflow (check outdated, categorize risk, test, deploy)
  • Automated update tools (Dependabot, Renovate, npm-check-updates)
  • Handling breaking changes and rollback plans

Monorepo Management

references/monorepo-patterns.md

Managing multiple related packages in a single repository with shared dependencies.

  • Workspace tools (pnpm, npm, yarn workspaces)
  • Monorepo structure and organization
  • Build optimization (Nx, Turborepo)
  • Versioning and publishing strategies

Transitive Dependencies

references/transitive-dependencies.md

Dealing with dependencies of your dependencies (indirect dependencies).

  • Viewing dependency trees (npm ls, pnpm why, pipdeptree)
  • Resolving transitive conflicts (overrides, resolutions, constraints)
  • Security risks and version conflicts
  • Best practices (use sparingly, document, test)

Ecosystem-Specific Guides

references/ecosystem-guides.md

Language and package-manager-specific best practices.

  • Node.js (npm, yarn, pnpm comparison and best practices)
  • Python (pip, poetry, conda)
  • Rust (cargo), Go (go mod), Java (maven, gradle)
  • PHP (composer), .NET (nuget)

Anti-Patterns

references/anti-patterns.md

Common mistakes to avoid when managing dependencies.

  • Critical anti-patterns (not committing lockfiles, wildcards, ignoring audits)
  • Dangerous anti-patterns (never updating, deprecated packages)
  • Moderate anti-patterns (overusing overrides, ignoring peer deps)

Navigation: Templates

Node.js

assets/nodejs/

Python

assets/python/

Automation

assets/automation/


Supply Chain Security

assets/automation/template-supply-chain-security.md — Production-grade dependency security.

Related templates:

Key Sections

  • SBOM Generation — CycloneDX, SPDX formats; CI/CD integration
  • AI BOM (Emerging) — Extended SBOM for AI-native systems (models, datasets, training artifacts)
  • Provenance & Attestation — SLSA levels, Sigstore signing, npm provenance
  • Vulnerability Management — Triage workflow, severity SLAs, scanning tools
  • Upgrade Playbooks — Batching strategy, rollback procedures
  • Pinning & Reproducibility — Lockfiles, hash pinning, version constraints
  • EU Cyber Resilience Act — SBOM requirements effective Dec 2027

Do / Avoid

GOOD: Do

  • Generate SBOM for every release
  • Sign release artifacts (Sigstore/cosign)
  • Run vulnerability scans in CI/CD
  • Fix critical vulnerabilities within 24 hours
  • Use lockfiles for reproducible builds
  • Verify npm package provenance
  • Batch non-security updates by risk level

BAD: Avoid

  • Publishing without SBOM
  • Using unsigned packages in production
  • Ignoring vulnerability scanner output
  • Updating all dependencies at once
  • Using wildcard version ranges (
    *
    ,
    >=
    )
  • Committing without updating lockfile
  • Bypassing security gates "just this once"

Anti-Patterns

Anti-PatternProblemFix
No SBOMCan't respond to supply chain attacksGenerate SBOM in CI/CD
Unsigned artifactsTampering undetectableSign with Sigstore
Floating versionsBuild not reproducibleUse lockfiles + exact versions
All-at-once updatesHard to bisect regressionsBatch by risk level
npm install in CINon-deterministicUse
npm ci
No audit gateVulnerabilities ship to prodGate deployments on audit

AI-Generated Dependency Risks

WARNING: AI coding agents can introduce vulnerable or non-existent packages at scale (Endor Labs, 2025).

The Problem

AI tools accelerate coding but introduce supply chain risks:

  • Hallucinated packages — AI suggests packages that don't exist (typosquatting vectors)
  • Vulnerable dependencies — AI recommends outdated or CVE-affected versions
  • Unnecessary dependencies — AI over-relies on packages for simple tasks

Best Practices

DoDon't
Treat AI-generated code as untrusted third-party inputBlindly accept AI dependency suggestions
Enforce same SAST/SCA scanning for AI-generated codeSkip security review for "AI-written" code
Verify all AI-suggested packages actually existTrust AI to know current package versions
Integrate security tools into AI workflows (MCP)Allow AI to add dependencies without review
Vet MCP servers as part of supply chainUse unvetted AI integrations

Validation Checklist

Before accepting AI-suggested dependencies:

  • Package exists on registry (npm, PyPI, crates.io)
  • Package name is spelled correctly (no typosquatting)
  • Version is current and maintained
  • npm audit
    /
    pip-audit
    shows no vulnerabilities
  • Weekly downloads >1000 (established package)
  • Last commit <6 months (actively maintained)

Optional: AI/Automation

Note: AI assists with triage but security decisions need human judgment.

  • Automated PR triage — Categorize dependency updates by risk
  • Changelog summarization — Summarize breaking changes in updates
  • Vulnerability correlation — Link CVEs to affected packages

Bounded Claims

  • AI cannot determine business risk acceptance
  • Automated fixes require security team review
  • Vulnerability severity context needs human validation

Quick Decision Matrix

ScenarioRecommendation
Adding new dependencyCheck Bundlephobia, npm audit, weekly downloads, last commit
Updating dependenciesUse
npm outdated
, update in batches, test in staging
Security vulnerability foundUse
npm audit fix
, review CHANGELOG, test, deploy immediately
Monorepo setupUse pnpm workspaces or Nx/Turborepo for build caching
Transitive conflictUse
overrides
sparingly, document why, test thoroughly
Choosing JS package managerpnpm (fastest, disk-efficient), Bun (7× faster), npm (most compatible)
Python environmentuv (10-100× faster), Poetry (mature), pip+venv (simple), conda (data science)

Core Principles

1. Always Commit Lockfiles

Lockfiles ensure reproducible builds across environments. Never add them to

.gitignore
.

Exception: Don't commit

Cargo.lock
for Rust libraries (only for applications).

2. Use Semantic Versioning

Use caret (

^
) for most dependencies, exact versions for mission-critical, avoid wildcards (
*
).

{
  "dependencies": {
    "express": "^4.18.0",  // Allows patches and minors
    "critical-lib": "1.2.3"  // Exact for critical
  }
}

3. Audit Dependencies Regularly

Run security audits weekly, fix critical vulnerabilities immediately.

npm audit
npm audit fix

4. Minimize Dependencies

The best dependency is the one you don't add. Ask: Can I implement this in <100 LOC?

5. Update Regularly

Update monthly or quarterly. Don't let technical debt accumulate.

npm outdated
npm update

6. Use Overrides Sparingly

Only override transitive dependencies for security patches or conflicts. Document why.

{
  "overrides": {
    "axios": "1.6.0"  // CVE-2023-xxxxx fix
  }
}

Related Skills

For complementary workflows and deeper dives:


External Resources

See

data/sources.json
for curated resources:

  • Package managers: npm, pnpm, Yarn, pip, Poetry, Cargo, Go modules, Maven, Composer
  • Semantic versioning: SemVer spec, version calculators, constraint references
  • Security tools: Snyk, Dependabot, GitHub Advanced Security, OWASP Dependency-Check, pip-audit, cargo-audit, Socket.dev, Renovate
  • Lockfile management: Official docs for package-lock.json, poetry.lock, Cargo.lock, pnpm-lock.yaml
  • Monorepo tools: pnpm workspaces, npm workspaces, Yarn workspaces, Nx, Turborepo, Lerna, Bazel
  • Analysis tools: Bundlephobia, npm-check-updates, depcheck, pipdeptree, cargo tree
  • Supply chain security: SLSA framework, SBOM (CISA), Sigstore, npm provenance, OpenSSF Scorecard
  • Best practices: npm/Poetry/Cargo guides, ACM Queue articles, dependency hell references
  • Version management: nvm, pyenv, rustup, asdf
  • Learning resources: npm guides, Python Packaging User Guide, Rust Book, Monorepo.tools

Usage Notes

For Codex / Claude Code:

  • Use this skill when users need dependency management guidance
  • Reference specific resources based on the task (lockfiles, security, updates)
  • Provide ecosystem-specific guidance (Node.js, Python, Rust)
  • Always recommend security audits and reproducible builds
  • Encourage minimal dependencies and regular updates
  • Link to templates for common configurations

Best Practices:

  • Always commit lockfiles (except Cargo.lock for libraries)
  • Use semantic versioning (caret for most deps, exact for critical)
  • Audit dependencies weekly (
    npm audit
    ,
    pip-audit
    ,
    cargo audit
    )
  • Update dependencies monthly or quarterly (not all at once)
  • Choose package manager based on project needs (pnpm for speed, Poetry for Python apps)
  • Document dependency choices in ADRs (Architecture Decision Records)

Success Criteria: Dependencies are minimal, well-maintained, secure, reproducible across environments, and regularly audited for vulnerabilities.


Trend Awareness Protocol

IMPORTANT: When users ask recommendation questions about package managers, dependency tools, or supply chain security, you MUST use WebSearch to check current trends before answering.

Trigger Conditions

  • "What's the best package manager for [ecosystem]?"
  • "What should I use for [dependency management/security]?"
  • "What's the latest in dependency management?"
  • "Current best practices for [npm/pnpm/Poetry]?"
  • "Is [tool/approach] still relevant in 2026?"
  • "[pnpm] vs [npm] vs [yarn]?"
  • "Best dependency security scanner?"

Required Searches

  1. Search:
    "dependency management best practices 2026"
  2. Search:
    "[specific tool] vs alternatives 2026"
  3. Search:
    "supply chain security trends January 2026"
  4. Search:
    "[package manager] features 2026"

What to Report

After searching, provide:

  • Current landscape: What dependency tools are popular NOW
  • Emerging trends: New package managers, security tools, or patterns gaining traction
  • Deprecated/declining: Tools/approaches losing relevance or support
  • Recommendation: Based on fresh data, not just static knowledge

Example Topics (verify with fresh search)

  • Package managers (pnpm, npm, yarn, Poetry, uv for Python)
  • Security scanning (Snyk, Dependabot, Socket.dev)
  • Supply chain security (SBOM, Sigstore, SLSA)
  • Monorepo tools (Nx, Turborepo, Bazel)
  • Lockfile and reproducibility patterns
  • Automated dependency updates (Renovate, Dependabot)