ultron

Core Ultron behavior skill. Governs default agent: tone matching, memory access, error handling, user personalization, and response formatting. Auto-active on every response. Triggers on any message that doesn't clearly belong to a specialist agent.

Coder agent skill. Deep code analysis, complex algorithms, large refactors, performance optimization, and thorough code review. For when builder isn't enough. Auto-triggers on: refactor, optimize, performance, review, algorithm, architecture.

Researcher agent skill. Web search, doc lookup, tech comparisons, synthesized answers. Multi-source verification, no hallucination. Auto-triggers on: research, find, what is, how does, best way, should I use, compare, latest, docs for.

Builder agent skill. Implements backend logic, APIs, data models, CLI tools, and full features using file tools and bash. Reads before writing, tests after writing. Auto-triggers on: build, create, implement, write, add, make, scaffold.

Debugger agent skill. Finds root causes of bugs, errors, and crashes. Reads code, reproduces errors, fixes at source — never patches symptoms. Auto-triggers on: fix, bug, error, crash, broken, not working, exception, failing, wrong output.

Designer agent skill. Creates UI components, HTML pages, CSS, and frontend code. Dark-theme first, minimal, responsive. Reads existing styles before writing new ones. Auto-triggers on: design, UI, layout, component, page, style, frontend.

Planner agent skill. Decomposes complex requests into ordered subtasks, assigns each to the correct specialist agent, estimates dependencies, and tracks completion. Auto-triggers when request involves multiple steps, systems, or agents.

Verifier agent skill. Runs tests, checks regressions, validates changes. Passes cleanly or escalates to debugger with exact failure details. Auto-triggers after any build/design/fix, or when user says: test, verify, check, does it work, validate.

Ultra-compressed commit message generator. Cuts noise from commit messages while preserving intent and reasoning. Conventional Commits format. Subject ≤50 chars, body only when "why" isn't obvious. Use when user says "write a commit", "commit message", "generate commit", "/commit", or invokes /caveman-commit. Auto-triggers when staging changes.

Quick-reference card for all caveman modes, skills, and commands. One-shot display, not a persistent mode. Trigger: /caveman-help, "caveman help", "what caveman commands", "how do I use caveman".

Ultra-compressed code review comments. Cuts noise from PR feedback while preserving the actionable signal. Each comment is one line: location, problem, fix. Use when user says "review this PR", "code review", "review the diff", "/review", or invokes /caveman-review. Auto-triggers when reviewing pull requests.

Ultra-compressed communication mode. Cuts token usage ~75% by speaking like caveman while keeping full technical accuracy. Supports intensity levels: lite, full (default), ultra, wenyan-lite, wenyan-full, wenyan-ultra. Use when user says "caveman mode", "talk like caveman", "use caveman", "less tokens", "be brief", or invokes /caveman. Also auto-triggers when token efficiency is requested.

Compress natural language memory files (CLAUDE.md, todos, preferences) into caveman format to save input tokens. Preserves all technical substance, code, URLs, and structure. Compressed version overwrites the original file. Human-readable backup saved as FILE.original.md. Trigger: /caveman:compress <filepath> or "compress memory file"