Claude Agents Pack

Lean Claude Code agent pack for fast setup, safer skill loading, and production-ready workflows.

Why this pack

• Fast onboarding: install in minutes with npx.
• Clear execution flow: plan, implement, review, and document with purpose-built agents.
• Safer defaults: on-demand skills + deny-by-default skill permissions.
• Operationally ready: built-in validation and release automation.

Quick jump: Agents · Skills Matrix · Skills · Full Docs

Quick Start

Requires: Node.js and npm

# Via npx (recommended)
npx @shahboura/agents-claude --global

# Alternative (direct npm install)
npm install -g @shahboura/agents-claude && agents-claude --global

# Project install (current directory only)
npx @shahboura/agents-claude --project .

# Install with specific languages only
npx @shahboura/agents-claude --global --languages python,typescript

# Update existing installation
npx @shahboura/agents-claude --update

# Force update both global + current project scopes
npx @shahboura/agents-claude --update --all

# Uninstall
npx @shahboura/agents-claude --uninstall

# Global uninstall
npx @shahboura/agents-claude --uninstall --global

# Uninstall both global + current project scopes
npx @shahboura/agents-claude --uninstall --all

# Check detected installation scopes
npx @shahboura/agents-claude --status

Install behavior note:

• npx/npm installs from the published npm package version (deterministic release artifact).

Package naming:

• Install from @shahboura/agents-claude (scoped package).
• After global install, run the CLI command as agents-claude.
• agents-claude is the installer/maintenance CLI; day-to-day use runs in claude.
• If you see unscoped agents-claude on npm, treat it as a different package.

Uninstall behavior:

• npx @shahboura/agents-claude --uninstall targets the current project by default.
• Use --global or --all for explicit scope control.
• Uninstall removes installer-managed files via install manifest tracking.
• If status shows installed (version-marker), uninstall removes marker files only. Run --update first, then uninstall for full managed-file cleanup.
• If manifest trust checks fail, uninstall removes only manifest/version markers and skips broad file deletion for safety.
• Project backups: <project>/.claude/.backups/<timestamp>--<operation>--<scope>/
• Global backups: ~/.claude/.backups/<timestamp>--<operation>--<scope>/
• Backup retention: latest 10 sessions and sessions newer than 30 days.

Restore from backup:

1. Open the latest backup session folder.
2. Review backup-manifest.json for file paths.
3. Copy files back to their original paths.

Update behavior:

• npx @shahboura/agents-claude --update auto-detects and updates installed scopes (global and/or current project).
• Use --all, --global, or --project [dir] to force explicit update scope.

Settings behavior:

• Installer merges only missing safe defaults in .claude/settings.json.
• Existing permissions, sandbox, and hooks are preserved and never overwritten.
• If CLAUDE.md already exists before install, installer leaves it unmanaged and will not remove it on uninstall.
• Installer uses an internal managed marker so only installer-managed CLAUDE.md is eligible for removal on uninstall.

Install safety behavior:

• Installer blocks writes to symlink destinations for managed files.
• Replace managed-path symlinks with regular files/paths before re-running install.

Scope behavior:

• --global targets ~/.claude only.
• --project [dir] targets <project>/.claude and installer-managed CLAUDE.md.
• --update auto-detects installed scopes and updates each.
• --uninstall defaults to current project scope only.
• --all applies update/uninstall to both global and current project scopes.

First run in Claude Code:

claude
/init
@orchestrator Build a REST API with JWT auth

Agents

| Agent | Best For | Allocated Skills (summary) | |-------|----------|----------------------------| | @orchestrator | Multi-phase coordination | Language skills + utility skills + blogger/brutal-critic | | @planner | Read-only architecture/planning | Language skills + utility skills | | @codebase | Feature implementation | Language skills + sql-migrations | | @review | Security/performance/code quality | Language skills + docs-validation + agent-diagnostics | | @docs | Documentation updates | docs-validation + project-bootstrap + agent-diagnostics | | @em-advisor | EM/leadership guidance | project-bootstrap + docs-validation + agent-diagnostics | | @blogger | Blog/video/podcast drafting | blogger + brutal-critic | | @brutal-critic | Final content quality gate | brutal-critic + blogger |

See full allowlists: Skills Matrix

Skill Loading (Claude)

• Skills live in .claude/skills/<name>/SKILL.md.
• Skills are loaded on demand via Claude Code skills.
• Use one relevant skill per task/phase by default; add another only for clear cross-domain work.
• If stack/domain is unclear, ask for clarification before loading.

Skill Scope Policy (Keep it)

• Yes, keep this policy. It keeps the pack focused and safe.
• Current scope is core-only skills (no optional skill packs).
• Additions should pass demand, clear-gap, ownership, and licensing/provenance checks.

Skill Permissions (Least Privilege)

Use least-privilege tools and selective skill invocation to prevent unrelated context/tool access.

This keeps skills focused by task and reduces accidental context bloat.

Skills

Type /skill-name in Claude Code to run:

| Skill | Description | |---------|-------------| | /api-docs | Generate API documentation | | /code-review | Comprehensive code review | | /generate-tests | Unit test generation | | /security-audit | Security audit | | /refactor-plan | Refactoring plan | | /create-readme | Generate README | | /architecture-decision | ADR creation | | /architecture-review | Architecture review | | /blog-post | Blog post creation | | /content-review | Content quality scoring | | /plan-project | Multi-phase project planning | | /1-on-1-prep | Meeting preparation |

Docs

• Getting Started
• Full Documentation