tachibot-mcp

Use when the user wants a security review of code or a diff — OWASP/CWE findings with severity, taint/data-flow analysis, and concrete fixes via security_review

Use when the user wants tests generated for code or a diff — enumerates edge cases and failure modes first, then emits runnable test code via testgen

Use when a task would benefit from a structured thinking technique but you're not sure which — recommends the right prompt technique for the intent (debug, spec, refactor, decision, design) and applies it

Use for decisions or judgments worth multiple independent perspectives — multi-model council with parallel analysis, search grounding, and synthesis (30-60s); for a quick single-call verdict use the jury tool directly

Use when the user shares or finishes a plan and wants it stress-tested before execution — adversarial multi-model pre-mortem with ranked risks, mitigations, and concrete plan edits

Use when code changes are ready for review — a diff exists, the user says "review my changes/PR", or a commit is about to happen. Multi-model diff review (Kimi, DeepSeek, GPT-5.5 panel + Gemini judge) with a MERGEABLE verdict

Use when TachiBot tools seem missing, a provider isn't working, or the user asks to set up/configure TachiBot or its API keys — runs doctor and walks through fixing the gaps

Use when a feature request is loose or ambiguous and needs a reviewable spec before planning — turns "add X somehow" into user stories, acceptance criteria, out-of-scope, and open questions via spec_writer

Use when the user asks what TachiBot can do, which tools/skills exist, or which API keys are configured — help and discovery entry point

Use when the user hits an error, exception, or stack trace and the cause isn't obvious — returns RANKED root-cause hypotheses with the cheapest discriminating check for each, via debug_triage

Algorithm analysis with parallel multi-model reasoning — deepseek_algo (lead), qwen_algo, qwen_reason, and minimax_code tackle your problem from different angles

Long-context analysis — examine a large document or codebase under a focused lens using active retrieval over Kimi's 256K window, surfacing insights, relations, and inconsistencies

Iterative reflexion loop — generate, critique against EXTERNAL evidence (real tests, search), then revise until it converges. Grounded self-correction, not sycophantic self-review

Tree-of-Thought reasoning — branch a problem into multiple solution paths, explore each, score and prune branches via a model jury, then synthesize the best. Non-linear reasoning that can backtrack

Multi-model implementation planning — creates bite-sized TDD plans using planner_maker's council pipeline with goal-oriented checkpoints

Decompose problems into sub-problems, then deep-dive each one with sequential multi-model reasoning

Break down complex problems before implementation using multi-model reasoning and prompt techniques

Multi-model reasoning with modes like deep-reasoning, architecture-debate, research, analyze

Chain-of-thought reasoning with optional multi-model execution using nextThought