Use when drafting a task issue, Git commit message, pull request description, or evidence-backed task, milestone, phase or final implementation report, review/revision campaign plan or report, package preparation result, curated Git release highlights, release notes or release outcome for an SDD project. Adapt the summary to feature, bug, code health, performance, security, testing, documentation, integration, or compatibility work without inventing results.
When implementation intent is clear — deliver the authorized outcome with a small correct diff, meaningful tests, and an uncommitted verification handoff.
Recommend and execute an efficient Astra, Sol, or Luna route with GPT-5.6 compatibility with low-through-max reasoning for Codex project work. Prefer bounded direct tool concurrency, but automatically create or reuse a model-specific leaf agent when route-fit benefit clearly exceeds startup and aggregation overhead; no extra user permission is required. Use for code changes, tests, reviews, routed implementation, model recommendations, usage queries, retuning, and requests to disable, exit, restore, or check this Skill for the current project. Use the legacy strict state machine only when the user explicitly requests strict ledger auditing. Never auto-select Ultra or create a new top-level Codex task.
Access atypica.ai's AI-powered business research capabilities through MCP protocol. Use when users need to conduct consumer research, market analysis, product testing, or generate insights about business questions. Supports creating research sessions, sending messages with AI execution, handling user interactions (question prompts and plan confirmations), retrieving conversation history, checking research status, accessing generated reports/podcasts, and searching AI personas for interviews.
Red team engagement planning is the foundational phase that defines scope, objectives, rules of engagement (ROE), threat model selection, and operational timelines before any offensive testing begins.
Use before pushing, force-pushing, marking ready for review, or claiming checks pass on a deepseek-harness branch, and immediately after gh stack sync publishes rewritten branches, to select the smallest tests and checks that cover the outgoing or just-published diff without reflexively running the full repository suite.
Use when drafting, rewriting, reviewing, or decomposing technical product requirements for product systems, micro-processors, sensors, actuators, or functions, diagnostics, interfaces, or specifications, engineering change requests. Helps turn engineering know-how, notes, standards, customer notes and test expectations from conversational casual writing into clear, atomic, measurable, traceable，unambiguous requirements with acceptance criteria.
Issue Planning and Automation prompt that generates comprehensive project plans with Epic > Feature > Story/Enabler > Test hierarchy, dependencies, priorities, and automated tracking.
Walk the user through reviewing a change: what it is for, what to look at closely, and how to test it. Use when the user says "walkthrough", "walk me through this change", or "human review"
Workflow 1: Full idea discovery pipeline. Orchestrates research-lit → idea-creator → novelty-check → research-review to go from a broad research direction to validated, pilot-tested ideas. Use when user says "找idea全流程", "idea discovery pipeline", "从零开始找方向", or wants the complete idea exploration workflow.
Deterministic play-testing over MCP — pause-and-step exact tick windows, the input latch lag, spawn placement, 3D pickup radius, what serializes, and Play on a duplicated world (Stop drops it).
Progressively activates todo tests one at a time, updates implementation code until each passes (verified via Wallaby), checks off matching design doc progress, then moves to the next—following the design doc as the single source of truth.
METAINFORMANT rules for directory src/metainformant/protein/database. Use when editing, adding tests, or reviewing code under this path. Read the linked AGENTS.md first; use uv only, write outputs to output/, real implementations.
Safety hooks for Claude Code — 916 pre-built hooks that prevent file deletion, credential leaks, git disasters, and token waste during autonomous AI coding sessions. 296 test files. Install with npx github:yurukusa/cc-safe-setup.