Claude Agentic Framework

Display configuration status and health information for framework agents

Perform comprehensive framework health check and analysis

Build exactly what a spec describes, then loop build ⇆ spec-compliance review autonomously until the review passes clean

Route and orchestrate a development objective end-to-end through specialist sub-agents

Display all available agents with capabilities and status

Use this agent to RUN command-line work and return distilled results, and to write or refactor shell scripts. Delegate to it for: executing builds, test suites, linters and validator scripts and reporting pass/fail with the exact failing lines; GitHub CLI work (gh auth/repo/pr/issue/run and gh api with --paginate and --jq); querying or transforming JSON and YAML with jq and yq; git inspection (status --porcelain, diff --stat, rev-list, merge-base, show, blame); HTTP probes with curl; log and output grinding where hundreds of lines must be reduced to one answer; and multi-step shell pipelines whose intermediate output the caller does not need. Also use it to author Bash/POSIX shell scripts: CI/CD and deployment automation, Linux/Unix administration, container and cloud-CLI scripting, error handling and signal trapping, portability across sh/bash/dash, and ShellCheck-clean refactoring. This agent executes commands itself and never delegates command-line work to another agent. <example>\nContext: The user wants the repo's validator run and the result explained.\nuser: "Run the consistency validator and tell me what is failing."\nassistant: "I'll use the bash-expert agent to run the validator and report the exit code plus only the failing checks."\n<commentary>\nRunning a command and reducing several hundred lines of output to the answer is exactly what bash-expert is for — the raw output never needs to reach this context.\n</commentary>\n</example>\n<example>\nContext: The user needs GitHub state summarized.\nuser: "Which open PRs have failing checks?"\nassistant: "I'll use the bash-expert agent to query the GitHub CLI and return just the PR numbers and the failing check names."\n<commentary>\ngh pr list and gh api with a --jq filter are bash-expert's toolkit; it returns a short table instead of paginated JSON.\n</commentary>\n</example>\n<example>\nContext: The user has a quoting bug in a shell script.\nuser: "My Bash script isn't handling filenames with spaces correctly"\nassistant: "Let me invoke the bash-expert agent to fix the quoting and word-splitting, then run the script to prove the fix."\n<commentary>\nQuoting and word-splitting are core Bash expertise, and bash-expert can execute the script to verify — authoring plus verification in one agent.\n</commentary>\n</example>

Use this agent when you need comprehensive code review for any code changes, implementations, or pull requests. This agent MUST be consulted before merging, deployment, or acceptance of any code. Invoke proactively for reviewing commits, analyzing new features, evaluating refactoring, checking security implications, assessing performance impact, verifying error handling, ensuring code conventions, validating test coverage, reviewing API designs, checking documentation, identifying technical debt, evaluating architectural decisions, spotting concurrency issues, detecting memory leaks, reviewing database queries, checking input validation, verifying accessibility, ensuring backward compatibility, reviewing configuration, and checking dependency updates.\n\n<example>\nContext: The user has just implemented a new authentication feature.\nuser: "I've implemented the new login functionality with JWT tokens"\nassistant: "I'll use the code-review-gatekeeper agent to perform a comprehensive review of your authentication implementation, checking for security vulnerabilities, proper token handling, and architectural alignment."\n<commentary>\nSince new authentication code has been written, use the Task tool to launch the code-review-gatekeeper agent to review it for security, correctness, and best practices.\n</commentary>\n</example>\n\n<example>\nContext: The user is about to merge a pull request.\nuser: "The feature is complete and all tests are passing. Ready to merge PR #234"\nassistant: "Before merging, let me invoke the code-review-gatekeeper agent to perform a final comprehensive review of all changes in PR #234."\n<commentary>\nBefore any merge, the code-review-gatekeeper must review all changes to catch potential issues.\n</commentary>\n</example>\n\n<example>\nContext: The user has refactored a complex algorithm.\nuser: "I've refactored the search algorithm to improve performance"\nassistant: "I'll have the code-review-gatekeeper agent analyze your refactoring to verify behavior preservation, performance improvements, and check for any introduced bugs or edge cases."\n<commentary>\nRefactoring requires careful review to ensure no regressions, so invoke the code-review-gatekeeper.\n</commentary>\n</example>

Use this agent when you need deep analysis, evaluation, or investigation of any complex information, systems, or problems. This includes examining code quality, assessing project completeness, decomposing requirements, identifying gaps and risks, evaluating technical solutions, diagnosing issues, comparing alternatives, or synthesizing findings from multiple sources. The agent excels at transforming raw information into structured insights and actionable intelligence.\n\nExamples:\n- <example>\n Context: User wants to understand the current state of their codebase and identify areas for improvement.\n user: "Can you analyze my project structure and identify any architectural issues or code quality problems?"\n assistant: "I'll use the comprehensive-analyst agent to perform a thorough analysis of your project."\n <commentary>\n Since the user is asking for analysis of their codebase, use the Task tool to launch the comprehensive-analyst agent to examine the project structure, code quality, and architectural patterns.\n </commentary>\n</example>\n- <example>\n Context: User needs to understand dependencies and risks in a complex system.\n user: "I need to understand all the dependencies in my microservices architecture and identify potential failure points"\n assistant: "Let me deploy the comprehensive-analyst agent to map out your system dependencies and identify risks."\n <commentary>\n The user needs complex system analysis, so use the comprehensive-analyst agent to trace dependencies and identify potential issues.\n </commentary>\n</example>\n- <example>\n Context: User wants to evaluate multiple technical solutions.\n user: "We're considering three different database solutions for our new project. Can you help evaluate them?"\n assistant: "I'll engage the comprehensive-analyst agent to compare and evaluate these database solutions for your project."\n <commentary>\n Since the user needs comparative analysis of technical solutions, use the comprehensive-analyst agent to evaluate options based on various criteria.\n </commentary>\n</example>

Use this agent when you need to write, refactor, analyze, or optimize any C# code. This includes creating new .NET applications, ASP.NET Core web services, WPF/WinForms/MAUI desktop applications, Unity game scripts, Azure cloud solutions, Entity Framework integrations, Blazor web apps, microservices, or console applications; fixing memory leaks and performance issues; implementing LINQ queries and async/await patterns; managing NuGet dependencies; working with .NET Core, .NET Framework, or .NET 8+; ensuring code follows C# best practices and SOLID principles. <example>\nContext: The user needs to implement a RESTful API with ASP.NET Core.\nuser: "I need to create an ASP.NET Core Web API for managing user authentication and authorization"\nassistant: "I'll use the csharp-expert agent to help you build a secure ASP.NET Core Web API with proper authentication and authorization."\n<commentary>\nSince this involves writing C# code for web API development with security requirements, the csharp-expert should be invoked.\n</commentary>\n</example>\n<example>\nContext: The user is encountering async/await deadlock issues in their C# code.\nuser: "My async method is causing a deadlock when I use .Result"\nassistant: "Let me invoke the csharp-expert agent to help resolve this async/await deadlock issue."\n<commentary>\nAsync/await deadlock issues are a common C# pattern that the csharp-expert specializes in.\n</commentary>\n</example>\n<example>\nContext: The user wants to refactor existing C# code to follow SOLID principles.\nuser: "Can you review this C# service class and refactor it to be more maintainable?"\nassistant: "I'll use the csharp-expert agent to review and refactor your C# code to follow SOLID principles and best practices."\n<commentary>\nRefactoring C# code to follow SOLID principles is a key responsibility of the csharp-expert.\n</commentary>\n</example>

Use this agent when you need to design database schemas, optimize queries, troubleshoot performance issues, implement data migrations, or architect data storage solutions. This includes designing relational schemas, creating indexes, writing complex SQL queries, optimizing query performance, implementing database migrations, choosing between SQL/NoSQL databases, designing data models, implementing replication/sharding, ensuring data integrity with constraints and transactions, optimizing backup/recovery strategies, and resolving deadlocks and performance bottlenecks. <example>\nContext: The user needs to design a database schema for a multi-tenant SaaS application.\nuser: "I need to design a PostgreSQL schema for a SaaS app with thousands of tenants - should I use row-level security or separate schemas?"\nassistant: "I'll use the database-specialist agent to design an optimal multi-tenant database architecture with proper isolation, performance, and scalability considerations."\n<commentary>\nSince this involves database schema design with multi-tenancy requirements, the database-specialist should be invoked.\n</commentary>\n</example>\n<example>\nContext: The user's queries are running slowly in production.\nuser: "My PostgreSQL query takes 30 seconds - it's a join across 5 tables with millions of rows"\nassistant: "Let me invoke the database-specialist agent to analyze and optimize this query with proper indexing, query rewriting, and execution plan analysis."\n<commentary>\nQuery optimization and performance tuning are core database concerns that the database-specialist specializes in.\n</commentary>\n</example>\n<example>\nContext: The user needs to choose between database technologies.\nuser: "Should I use PostgreSQL, MongoDB, or Cassandra for my time-series IoT data with 100k writes/second?"\nassistant: "I'll use the database-specialist agent to evaluate database options based on your write-heavy time-series requirements and recommend the optimal solution."\n<commentary>\nDatabase selection and architecture decisions are key responsibilities of the database-specialist.\n</commentary>\n</example>

Diagnose and fix agent routing, loading, and configuration issues by inspecting agents/*.md frontmatter, the claude.json registry, hook registration, and validator output — use when an agent is "not found", tasks route to the wrong specialist, models mismatch, hooks do not fire, or validate-consistency.sh fails.

Recommends which framework agent (or multi-agent workflow) should handle a task; use when the user asks which agent to use, how to delegate a piece of work, or how to sequence agents for a multi-domain task.

Generate idiomatic project scaffolding, config files, and boilerplate per language — use when starting a new project, adding a service, or bootstrapping a library/package (e.g. "scaffold a Rust CLI", "set up a new FastAPI project").

Iteratively improve a code change by having the matching specialist agent score the diff 0-100 against an explicit rubric, rewriting the weakest parts, and repeating until the score plateaus. Use after the quality bar is green (build/lint/format/tests clean) but before the review gates — e.g. polishing an implementation or refactor, or when asked to "clean up" code without concrete criteria. Code-side sibling of self-scoring-loop.

Verify the framework's toolchain (git, bash, jq, gh, yq, PowerShell 7+, Node.js/npx, uv/uvx, shellcheck, Claude Code CLI) is installed and working — use when setting up the framework, after system updates, or when hooks, validators, MCP servers, or the executor agents fail with "command not found".

5 hooks across 5 events