checkpoint-guardian
fatih-developerAutomatic risk assessment before every critical action in agentic workflows. Detects irreversible operations (file deletion, database writes, deployments, payments), classifies risk level, and requires confirmation before proceeding. Triggers on destructive keywords like deploy, delete, send, publish, update database, process payment.
multi-brain
fatih-developerEvaluate complex requests from 3 independent perspectives (Creative, Pragmatic, Comprehensive), reach consensus, then produce complete outputs. Use for architecture decisions, creative content, analysis, and any task where multiple valid approaches exist.
task-decomposer
fatih-developerBreak down large, complex, or ambiguous tasks into independent subtasks with dependency maps, execution order, and success criteria. Plan first, then execute step by step. Triggers on "how should I do this", "where do I start", "plan the project", "break it down", "implement" or whenever a task involves multiple phases.
agent-reviewer
fatih-developerAfter an agentic task completes, perform a retrospective analysis across 6 dimensions (goal alignment, efficiency, decision quality, error handling, communication, reusability). Score performance, identify inefficiency patterns, evaluate skill usage, and produce actionable improvement recommendations. Triggers on "how did it go", "retrospective", "review performance", "what could be better", or after any long agentic task completes.
react-flow
fatih-developerAnalyze, repair, migrate, and scaffold @xyflow/react codebases. Use when users ask to debug React Flow behavior, fix node/edge state wiring, improve type safety or performance, upgrade legacy React Flow APIs, preserve persisted graph compatibility, or generate a complete React Flow starter from scratch.
Parallel Planner
fatih-developerAnalyze multi-step tasks to identify which steps can run in parallel, build dependency graphs, detect conflicts (write-write, read-write, resource contention), and produce optimized execution plans. Triggers on 3+ independent steps, "speed up", "run simultaneously", "parallelize", "optimize" or any task where sequential execution wastes time.
tool-selector
fatih-developerBefore executing a task, analyze the available tool set (web search, code execution, file read/write, API calls, database queries, memory tools), select the optimal tools, plan execution order, and prevent unnecessary tool calls. Triggers on multi-step tasks, "which tool should I use", "plan the tools", or whenever multiple tools could apply.
output-critic
fatih-developerEvaluate every produced output (code, report, plan, data, API response) against type-specific quality criteria, score 1-10, make accept/reject decisions, and provide actionable improvement suggestions. Triggers on "evaluate", "check", "review", "quality control", "is this good enough", "score it", or before passing output to the next step in an agentic workflow.
multi-brain-experts
fatih-developerReplace generic perspectives with domain-specific expert roles selected dynamically per request. Automatically picks the 3 most relevant experts from a role pool (Security, Performance, UX, Cost, DX, Architecture, etc.) based on the task context.
error-recovery
fatih-developerWhen a step fails during an agentic task, classify the error (transient, configuration, logic, or permanent), apply the right recovery strategy, and escalate to the user when all strategies are exhausted. Triggers on error messages, exceptions, tracebacks, "failed", "not working", "retry", or when 2 consecutive steps fail.
project-analyzer
fatih-developerDeeply analyzes an existing software project and generates 3-4 comprehensive reports in the docs/analyze/ folder. Trigger when phrases like 'analyze the project', 'do code analysis', 'project review', 'architecture report', 'technical debt', 'project evaluation', 'codebase analysis' occur. Works when a project directory or repo URL is provided. If API endpoints are detected, it automatically generates a 4th report.
project-analyzer
fatih-developerAnalyzes an existing software project in depth and produces 3–4 comprehensive reports in the docs/analyze/ folder. Trigger when phrases like 'analyze the project', 'perform code analysis', 'project review', 'architecture report', 'technical debt', 'project evaluation', or 'codebase analysis' are mentioned. Operates when given a project directory or a repository URL. If API endpoints are detected, it automatically generates a 4th report.
Multi-Brain Debate
fatih-developerTwo-round debate protocol in which distinct perspectives challenge one another before reaching consensus. Round 1 presents independent opening positions; Round 2 allows targeted counter-arguments and concise rebuttals. Produces battle-tested decisions for high-stakes choices.
assumption-checker
fatih-developerBefore starting a task or taking a critical step, surface and verify the assumptions the agent is making. Checks 4 types - technical (libraries, APIs), data (files, formats), business logic (rules, scope), and user intent (what the user actually wants). Triggers on ambiguous requests, multi-step tasks, or whenever "are you sure", "check first", "don't assume" appears.
Multi-Brain Score
fatih-developerConfidence scoring overlay for multi-brain decisions. Each perspective rates its own confidence (1–10) with justification. Consensus uses scores as weights, flags low-confidence areas, and surfaces uncertainty explicitly.
memory-ledger
fatih-developerMaintain a structured ledger of decisions, discovered bugs and fixes, user preferences, constraints, current status, and failed approaches throughout multi-step agentic tasks. Auto-update after every significant step. Triggers on "where were we", "continue", "summarize status", "remember", or when a new agent instance takes over a task.
Context Compressor
fatih-developerCompress long conversation histories, large code files, research results, and documents by 70% without losing critical information. Triggers when context window fills up, when summarizing previous steps in multi-step tasks, before loading large files into context, or on "summarize", "compress", "reduce context", "save tokens".