dreamcode

DEEP multi-step research harness. Decomposes queries into sub-questions, executes parallel web searches via Pieces MCP, iteratively refines with gap detection, and synthesizes cited reports. Architecture inspired by GPT Researcher (27.6k stars), Vane/Perplexica (35.2k stars), and Gemini Deep Research. Use when user asks for deep research, investigation, analysis, comparison, or any task requiring exhaustive web research.

DEFAULT THINKING MODE — not a feature, but how the agent operates. Every task gets innovation-overdrive thinking: research, ground, reflect on contradictions, propose, then build. This is the agent's identity, not a command. Integrates MiMo-Code's 6-phase memory consolidation.

Chain execution manager: validates dependencies, enforces order, tracks violations. Use to verify that all required skill chains were executed and to get the correct execution order.

API design patterns, REST conventions, and endpoint standards. Use when creating or modifying API endpoints. Covers routing, request/response patterns, error handling, and versioning.

Architectural design, system patterns, and dependency management. Use when designing new modules, refactoring existing systems, or making architectural decisions. Covers layering, coupling, cohesion, and design patterns.

Trigger-driven automation definitions and runner management. Use when creating, running, or managing skill-based automation pipelines.

DEFAULT THINKING MODE — not a feature, but how the agent operates. Every task gets innovation-overdrive thinking: research, ground, reflect on contradictions, propose, then build. This is the agent's identity, not a command. Integrates MiMo-Code's 6-phase memory consolidation.

Customize opencode configuration, theme, keybindings, and agent settings. Use when modifying dreamcode/opencode behavior, adding custom agents, or adjusting system preferences.

Data science and statistical analysis best practices. Use for data analysis, statistical modeling, visualization, and numerical computation. Covers pandas/numpy patterns, statistical methodology, and reproducibility.

Docker, CI/CD, deployment, and infrastructure management. Use for container builds, CI pipeline changes, deployment configurations, and environment setup.

Documentation standards for code, APIs, and project docs. Use when writing docstrings, README files, API docs, or any project documentation.

Work with Effect v4 / effect-smol TypeScript code in this repo

META-SKILL — mandatory entry point for EVERY non-trivial prompt. Decomposes prompt into 5 orthogonal cursors (Temporal, Source, Gesture, Topic, People), fires parallel Pieces LTM searches with pagination until confidence >= threshold, then orchestrates the full neuro (10 iter) -> code-hardener (5 iter) -> implementation -> lint-fixer (5 loop) chain. Self-evolutionary: after every run, analyzes outcomes and updates its own structure. Integrates, wraps, and supersedes neuro, code-hardener, and lint-fixer.

Frontend development standards for React, TailwindCSS, and Vite. Use for UI components, pages, styling, and frontend architecture. Covers component patterns, state management, and performance.

Git operations, branching strategy, commit conventions, and PR workflow. Use for all git operations. Enforces clean history and conventional commits.

Mandatory post-implementation lint and type-checking skill that runs after ANY code change, bug fix, new feature, refactor, API change, or configuration modification. Ensures all TypeScript type errors, Effect-TS ecosystem issues, and lint errors are resolved. Use after every implementation that modifies source files.

Project onboarding and orientation. Use when first exploring the codebase, setting up the development environment, or understanding the project architecture.

Performance analysis, profiling, and optimization. Use when optimizing slow code, reducing memory, or scaling to larger datasets. Covers profiling, bottleneck identification, and optimization patterns.

Product-oriented thinking for understanding user needs, prioritizing features, and designing solutions. Use when the task involves product decisions, feature prioritization, or understanding user impact.

Python development standards, typing, imports, and project structure. Use for all Python code in the project. Covers modern Python tooling with ruff, mypy, pytest.

Quantum computing POC standards for QAE, QAOA, and hybrid quantum-classical algorithms. Use for quantum circuit design, simulator benchmarking, and honest reporting of quantum results.

React development standards for hooks, components, state management, and performance. Use for all React/JSX code.

Security review and vulnerability analysis. Use when handling sensitive data, authentication, authorization, input validation, or any security-relevant code. Based on OWASP Top 10 patterns.

Anti-slop frontend skill — reads the brief, infers design direction, ships interfaces that do not look templated. Real design systems when applicable, audit-first on redesigns, strict pre-flight check. Covers typography, layout, motion, accessibility, dark mode, and premium design taste. Ported from leonxlnx/taste-skill (53k★ GitHub).

META-SKILL — mandatory entry point for EVERY non-trivial prompt. Decomposes prompt into 5 orthogonal cursors (Temporal, Source, Gesture, Topic, People), fires parallel Pieces LTM searches with pagination until confidence >= threshold, then orchestrates the full neuro (10 iter) -> code-hardener (5 iter) -> implementation -> lint-fixer (5 loop) chain. Self-evolutionary: after every run, analyzes outcomes and updates its own structure. Integrates, wraps, and supersedes neuro, code-hardener, and lint-fixer.

API design patterns, REST conventions, and endpoint standards. Use when creating or modifying API endpoints. Covers routing, request/response patterns, error handling, and versioning.

Mandatory second-stage logic filter and architectural hardening skill that runs after NEURO for ANY code change, bug fix, new feature, refactor, API change, data contract change, performance optimization, security fix, architectural decision, configuration change, test modification, integration work, debugging, analysis, planning, or implementation. Executes exactly 5 mandatory iterations. Validates NEURO recommendations against repo truth, calls NEURO with filtered critique, then emits the only implementation plan opencode may follow. Use after any neuro skill usage. Use when code is about to be edited. Use for every non-trivial code modification.

Communication standards for explaining technical concepts to different audiences. Use when presenting results, writing explanations, or preparing presentations.

Data science and statistical analysis best practices. Use for data analysis, statistical modeling, visualization, and numerical computation. Covers pandas/numpy patterns, statistical methodology, and reproducibility.

Systematic debugging methodology. Use when encountering unexpected behavior, test failures, or production issues. Covers reproduce-isolate-fix-verify cycle.

Git operations, branching strategy, commit conventions, and PR workflow. Use for all git operations. Enforces clean history and conventional commits.

Codex-inspired Guardian AI — NEURO-powered safety supervisor that reviews agent actions on EVERY prompt before execution. Uses NEURO API as its brain. Validates code changes, catches security issues, prevents destructive operations. MANDATORY — cannot be skipped.

Mandatory post-implementation lint and type-checking skill that runs after ANY code change, bug fix, new feature, refactor, API change, or configuration modification. Ensures all ruff lint errors, mypy type errors, and ESLint issues are resolved. Use after every implementation that modifies source files. Use when code has been edited. Use for every response where file edits were made.

Mandatory external architectural and code-review harness for ANY non-trivial task including code changes, bug fixes, new features, refactors, API changes, data contract changes, performance optimizations, security fixes, architectural decisions, configuration changes, test modifications, integration work, debugging, analysis, planning, and implementation. Use when the user requests any modification, analysis, debugging, testing, integration, or planning work. Use for EVERY prompt that involves understanding or changing code, data, or configuration. When in doubt, ALWAYS use this skill. Skip only for trivial changes: typo fixes, formatting-only changes, one-line lint fixes, comment cleanup, simple unused import removal.

Systematic project planning and task decomposition. Use when starting a new feature, refactoring, debugging, or any multi-step task. Provides structured thinking frameworks, spec-first analysis, and staged implementation plans.

Product-oriented thinking for understanding user needs, prioritizing features, and designing solutions. Use when the task involves product decisions, feature prioritization, or understanding user impact.

Python development standards, typing, imports, and project structure. Use for all Python code in the project. Covers modern Python tooling with ruff, mypy, pytest.

Code quality enforcement, linting, type checking, and best practices. Use after every code change to ensure production-grade quality. Integrates with ruff, mypy, pylint, and other quality tools.

Quantum computing POC standards for QAE, QAOA, and hybrid quantum-classical algorithms. Use for quantum circuit design, simulator benchmarking, and honest reporting of quantum results.

Safe refactoring methodology. Use when restructuring existing code without changing behavior. Covers patterns for incremental improvement, testing during refactors, and risk management.