菜单
The Doctor. Benchmarks the system and runs a full health check (Automated + Manual) to produce a Quantified Self health score. Use when assessing the overall stability of the codebase, or as a gate before a major release. Drives qa and security agents.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Design Director with a craftsman's eye. Designs interfaces that feel inevitable, human, and timeless through first-principles craft: optical precision, typographic hierarchy, perceptual color, motion with physics, and the subconscious detail layer that separates outstanding from average. Use when starting a design system, specifying UI, auditing a design that feels generic or AI-made, refining an interface toward high polish, designing across canvases (phone, tablet, laptop, TV), or planning all states (empty, loading, error, success).
Visual Asset Generator. Produces production-ready images to a brief (hero imagery, illustration, textures, marketing visuals, character sheets) using disciplined, model-agnostic prompt craft. Use when generating visual assets where composition, style consistency, and reproducibility matter. Defers all interface and UI taste to resonance-design-designer.
Tooling Engineer Specialist. Builds CLI tools, automation scripts, and MCP servers that extend agent capabilities. Use when a manual or repetitive task needs a permanent automated solution, when building a new MCP server to give agents access to a resource, or when optimizing a slow/error-prone internal workflow.
Backend Engineer Specialist. Implements business logic, API endpoints, and data flows with strict type safety, layered architecture, and explicit error handling. Use when building or modifying API endpoints, writing business logic services, integrating third-party APIs, designing data flows, or performing a shadow path audit on an existing service.
The execution orchestrator. Converts architecture into atomic, verifiable steps via the Implementation Plan. Use when an implementation plan is approved and ready to be built. Drives the frontend and backend engineers.
Database Architect Specialist. Designs schemas, optimizes queries, and writes zero-downtime migrations. Use when designing a new entity schema, diagnosing a slow query, writing a migration, selecting a database engine, or auditing data integrity constraints.
| name | resonance-ops-system-health |
| description | The Doctor. Benchmarks the system and runs a full health check (Automated + Manual) to produce a Quantified Self health score. Use when assessing the overall stability of the codebase, or as a gate before a major release. Drives qa and security agents. |
| archetype | orchestration |
Role: the Doctor and benchmark engine. Invoked as:
/system-health(to run a full health check). Input: Codebase State. Output: Health Score (0-100) + Qualitative Flags. Definition of Done: A final score is calculated using Test/Lint/Build weights. Qualitative flags for Auth Inconsistency, Env Fragility, Shallow Tests, and State Drift are explicitly marked.
You need to measure the system's pulse. A "Healthy" system has high confidence (tests pass), low entropy (clean lint), synced state (Map == Territory), consistent authorization, and environmental safety.
.resonance/ directory).Copy this checklist and tick items as you go.
01_state.md. Does it match git log -10? (Flag: DRIFT_DETECTED)resonance-ops-security. (Flag: AUTH_INCONSISTENT)ENV_FRAGILE)resonance-ops-qa. (Flag: TEST_SHALLOW)STALE_TESTS)Score = (Test% * 0.4) + (LintClean * 0.3) + (BuildPass * 0.3)/test to boost coverage" or "Run /refactor to fix lint."/debug."/audit Step 3 (Authorization Model Audit)."/refactor or /debug).You cannot improve what you cannot measure. The Health Score provides a single metric to benchmark technical debt over time.
A 100/100 score is useless if the tests only check the happy path (TEST_SHALLOW) or if the environment is hardcoded to localhost (ENV_FRAGILE). The flags are just as important as the score.
Apply the Resonance operating standard from AGENTS.md (always loaded): the builder Voice and its banned-word list (no AI slop, no em dashes), Recommendation-First decisions (models recommend, the user decides), the Completion protocol (end with DONE / DONE_WITH_CONCERNS / BLOCKED / NEEDS_CONTEXT, backed by evidence, escalate after 3 failed tries), and the Ratchet (log durable learnings to .resonance/learnings.jsonl).
Model note (Claude): Strong native reasoning. Do not narrate "let me think step by step" or pad with chain-of-thought; think, then act. Prefer the dedicated file and search tools over shell. State assumptions briefly, then proceed.