MQjehovah/agent

monthly-business-review

月度经营复盘。适用于"XX月经营情况 / 经营分析 / 经营报告 / 月度复盘"类任务。固化已验证可用的批量 SQL 查询集与报告模板，一次并行查完、直接出报告，禁止自由探索数据库。

product-business-analysis

产品经营分析。分析销售、采购、库存、成本、现金流、生产等业务数据并生成报告。已按数字中台真实数据库校准：明确标注哪些指标数据库能算、哪些无数据源不得编造，每个维度附可用 SQL。

prove-it-tdd

严格 Prove-It 测试驱动开发模式。先写失败测试，证明它失败，再写代码，证明它通过，最后重构。

ship

Release engineer checklist. Sync main, run tests, audit coverage, push, open PR. Bootstraps test frameworks if missing. Use when "ready to ship", "deploy", "publish", "release this".

office-hours

YC-style product interrogation. Six forcing questions that expose demand reality, status quo, desperate specificity, narrowest wedge, observation, and future-fit. Use before any planning or coding. Triggers when user describes a new product idea, asks "is…

plan-ceo-review

CEO/founder-mode plan review. Four modes: EXPANSION (dream big), SELECTIVE EXPANSION (hold scope + cherry-pick), HOLD SCOPE (maximum rigor), SCOPE REDUCTION (strip to essentials). Use when user asks "think bigger", "strategy review", "rethink this", or…

review

Pre-landing PR review. Analyzes diff for SQL safety, race conditions, LLM trust boundary violations, and other structural issues. Confidence calibration (1-10) on every finding. Triggers: "review this", "code review", "check my code".

cso

Chief Security Officer audit. OWASP Top 10 + STRIDE threat model. Zero-noise: only report findings with confidence ≥8/10. Each finding includes concrete exploit scenario. Triggers: "security audit", "security review", "is this secure", "check for…

api-and-interface-design

Contract-first interface design. Use when designing APIs, module boundaries, or public interfaces.

browser-testing-with-devtools

Chrome DevTools for live runtime data. Use when building or debugging anything that runs in a browser.

ci-cd-and-automation

Shift Left CI/CD with quality gates. Use when setting up or modifying build and deploy pipelines.

code-review-and-quality

Multi-axis code review. Use before merging any change.

code-simplification

Reduce complexity while preserving behavior. Use when code works but is harder to read or maintain than it should be.

context-engineering

Feeds agents the right information at the right time. Use when starting a session, switching tasks, or output quality drops.

debugging-and-error-recovery

Five-step triage for any failure. Use when tests fail, builds break, or behavior is unexpected.

deprecation-and-migration

Code-as-liability mindset for deprecation. Use when removing old systems, migrating users, or sunsetting features.

documentation-and-adrs

Architecture Decision Records and API docs. Use when making architectural decisions, changing APIs, or shipping features.

doubt-driven-development

Adversarial fresh-context review of every non-trivial decision. Use when stakes are high (production, security, irreversible).

frontend-ui-engineering

Component architecture and UI development patterns. Use when building or modifying user-facing interfaces.

git-workflow-and-versioning

Trunk-based development with atomic commits. Use when making any code change (always).

idea-refine

Structured divergent/convergent thinking to turn vague ideas into concrete proposals. Use when you have a rough concept that needs exploration.

incremental-implementation

Build in thin vertical slices. Use when any change touches more than one file.

interview-me

One-question-at-a-time interview that extracts what the user actually wants. Use when requirements are vague, the ask is underspecified.

observability-and-instrumentation

Structured logging, metrics, and tracing. Use when adding telemetry or shipping anything that runs in production.

performance-optimization

Measure-first performance approach. Use when performance requirements exist or you suspect regressions.

planning-and-task-breakdown

Decomposes specs into small, verifiable tasks with acceptance criteria. Use when you have a spec and need implementable units.

security-and-hardening

OWASP Top 10 prevention and security hardening. Use when handling user input, auth, data storage, or external integrations.

shipping-and-launch

Pre-launch checklist and staged rollout. Use when preparing to deploy to production.

spec-driven-development

Writes a spec/PRD before any code. Use when starting a new feature, project, or significant change. Use when requirements are unclear.

test-driven-development

Drives development with tests. Use when implementing any logic, fixing any bug, or changing any behavior.

using-agent-skills

Maps incoming work to the right skill workflow. Use when starting a session or deciding which skill applies.

device-pull-bag

通过运维平台 API 拉取设备录包文件，上传云端并返回下载链接

device-remote-operations

通用设备远程运维技能，处理定位丢失、故障诊断、状态异常等常见设备故障

device-collision-handling

处理设备碰撞故障，包括碰撞检测、紧急停止、脱困和恢复

device-cannot-back-station

处理"无法返回工作站"或"前往充电点失败"故障

device-offline-recovery

处理设备离线（connect=false）或无响应的恢复流程

gitlab-merge

通过 GitLab API 将指定项目的源分支合入目标分支。全程云端操作，不克隆仓库。接收参数：<project_path> <source_branch> <target_branch>

code-review

三阶段结构化代码审查（发现→验证→分级），输出P0/P1/P2/历史遗留问题清单

plan-ceo-review

CEO/founder-mode plan review. Four modes: EXPANSION (dream big), SELECTIVE EXPANSION (hold scope + cherry-pick), HOLD SCOPE (maximum rigor), SCOPE REDUCTION (strip to essentials). Use when user asks "think bigger", "strategy review", "rethink this", or…

deploy

你有公司业务系统rosiwit-cloud的部署能力