70 commands.
10 agents.
One plugin.

Slash-command prompts for Claude Code. From project setup to deployment, with specialized agents and a static validator.

Local bun run validate output: 70 of 70 commands pass; quality grade A, 90.4 out of 100.
bun run validate70 / 70 passing

Install

As a Claude Code plugin

Run these inside Claude Code, then restart it.

/plugin marketplace add ursisterbtw/ccprompts
/plugin install ccprompts@ccprompts-marketplace

After restarting, enter a command such as /analyze-project in Claude Code.

From a clone

Needs Node.js 20 to 25 and Bun 1.1.34. Installs dependencies and runs every quality gate.

git clone https://github.com/ursisterbtw/ccprompts.git
cd ccprompts
bun install --frozen-lockfile
bun run ci

Command catalog

Twelve phases from initial workflow to enterprise scale, plus five task workflows. Filter by phase or search by name.

Showing 70 of 70

Claude Code commands by phase, command name, and description
PhaseCommandWhat it does
Initial workflow/analyze-projectDetects project characteristics and suggests relevant prompts and improvements.
Initial workflow/intelligent-chainChains commands from a plain-language workflow description.
Project setup/documentDocumentation generation and knowledge management prompts.
Project setup/learnPersonalized learning paths with exercises for Claude Code users.
Project setup/mcpMCP server configuration and testing prompts.
Development/backupDetects and backs up code, databases, configurations and dependencies.
Development/debug-sessionTroubleshoots Claude Code issues and session problems.
Development/migrateDatabase, API and infrastructure migrations with rollback and zero-downtime strategies.
Development/optimizeOptimization workflows for performance, security, maintainability and cost.
Development/refactorMulti-file refactoring with automated testing and rollback.
Security/audit-securitySecurity and quality audit prompts.
Security/complyCompliance automation and security hardening prompts.
Security/hardenEnd-to-end security hardening with defense in depth.
Security/incident-responseStructured response workflows for production incidents.
Testing/testTest generation and automation with coverage analysis.
Testing/troubleshootSystematic debugging help for software development issues.
Deployment/deployCI/CD pipeline and Infrastructure as Code prompts.
Deployment/gitGit workflow automation and repository management prompts.
Deployment/pre-commitPre-commit checks and quality gates.
Deployment/setup-ciCI/CD pipelines with security, testing and deployment across platforms.
Collaboration/code-reviewAutomated code review with analysis and feedback.
Collaboration/daily-standupStandup preparation with progress tracking and tool integration.
Collaboration/monitorObservability setup: metrics, logs, traces and alerts.
Collaboration/tech-debtTechnical debt detection, prioritization and resolution tracking.
Utilities/best-practicesBest practices analysis and enforcement for development workflows.
Utilities/knowledge-baseCreates and organizes organizational knowledge across formats.
Utilities/quick-fixTargeted fixes for common development issues.
Utilities/release-notesRelease note generation formatted for multiple audiences.
Utilities/smart-suggestPrompt recommendations based on project state and context.
Utilities/sprint-planningSprint planning with capacity management and goal setting.
Utilities/validate-environmentChecks that the environment can run Claude Code prompts.
Extras/health-checkProject health assessment across multiple dimensions.
Extras/modernizeLegacy modernization workflow chaining optimization prompts.
Extras/new-featureFeature workflow from planning to deployment.
Extras/workflow-builderInteractive interface for building prompt chains.
Agentic capabilities/agent-communicateInter-agent communication protocols for coordinating specialized agents.
Agentic capabilities/agent-learnAgent learning and adaptation from project patterns and outcomes.
Agentic capabilities/agent-monitorObservability and optimization for multi-agent systems.
Agentic capabilities/agent-orchestrateCoordinates multiple AI agents on complex tasks.
Agentic capabilities/agent-specializeRole specialization and capability assignment for agents.
Agentic capabilities/context-managerContext management using semantic understanding and knowledge graphs.
Agentic capabilities/context-persistSession context persistence and retention.
Agentic capabilities/mcp-configureMCP server configuration, management and integration.
Agentic capabilities/mcp-discoverDiscovers, evaluates and installs MCP servers.
Agentic capabilities/mcp-extendBuilds custom MCP servers and tools.
Agentic capabilities/workflow-automateProcess automation and multi-agent workflow design.
Agentic capabilities/workflow-visualVisual workflow design for agent orchestration.
AI-native development/ai-debugAI-assisted debugging, error detection and resolution.
AI-native development/ai-mentorPersonalized mentoring and skill development.
AI-native development/ai-pair-programPair programming with deep context awareness.
AI-native development/code-explainCode explanation and documentation.
AI-native development/code-generateGenerates projects and multi-file structures from templates and context.
AI-native development/pattern-detectDetects design patterns, architectural patterns and anti-patterns.
AI-native development/predictive-devPredictive analytics for software development.
AI-native development/refactor-semanticSemantics-preserving code transformation.
AI-native development/semantic-understandProgram analysis and semantic code understanding.
AI-native development/test-intelligentAI-assisted test generation and optimization.
Enterprise scale/analytics-advancedAnalytics dashboards and reporting for business intelligence.
Enterprise scale/compliance-enterpriseCompliance automation for frameworks such as SOC2, GDPR and HIPAA.
Enterprise scale/governancePolicy management, compliance automation and risk management.
Enterprise scale/knowledge-orgOrganizational knowledge systems and information architecture.
Enterprise scale/multi-repoCoordination across multiple repositories.
Enterprise scale/resource-manageResource optimization and capacity planning.
Enterprise scale/scale-optimizePerformance optimization for large-scale systems.
Enterprise scale/team-coordinateLarge-scale team coordination and cross-functional collaboration.
Tasks/1_explore_plan_codeStrict phased explore, plan and code workflow.
Tasks/2_continue_explore_plan_codeResumes that workflow after context loss from PLANNING.md and TASKS.md.
Tasks/3_sync_plan_tasksSyncs plan and task files with the current repository state.
Tasks/agent_initExplores the repository and generates tailored Claude Code agents.
Tasks/cursor_rules_generatorExplores the repository and generates .cursor/rules files.

Agents

agent-template-wizard
Creates new agents from the subagent template.
bash-shell-scripting
Bash scripts, deployment automation and shell pipelines.
documentation-writer
Technical, API and architectural documentation.
fastapi-optimizer
FastAPI async programming, performance and OpenAPI docs.
golang-pro
Go development with goroutines and channels.
javascript-expert
Developing, optimizing and debugging JavaScript applications.
performance-optimizer
System-wide performance and efficiency work.
python-pro
Advanced Python: decorators, generators and async.
rust-expert
Rust ownership, lifetimes, unsafe code and performance.
systems-architect
Real-time and ultra-low-latency system design.

Validation

Every command file is validated

Run bun run validate. It enforces exactly 70 command files and runs static safety analysis over each one.

Files passing
70 of 70
Error rate
0.0%
Warning rate
37.1%
Security score
100%
Quality grade
A, 90.4 of 100
Validation time
21 ms

Figures from the light-theme capture, not a live status. The dark-theme capture shows a separate run. These are static-analysis results; prompt content is not executed. Read the safety model.

Local bun run validate output: 70 of 70 commands pass; quality grade A, 90.4 out of 100.

Safety model

Static checks, clearly scoped.

The automated validator performs static analysis. Its dry-run check confirms the command text can be passed to the wrapper as one literal argument. It does not execute prompt content.

A separate runtime boundary.

To run a command by hand, use the Dagger runner. The project is copied into an ephemeral container as a writable snapshot, and changes are not exported to the host. Dagger must be installed.

./scripts/safe-run.sh "npm test"

Read SAFETY.md in the repository before running untrusted commands.