
代码逆向分析知识库。提供多框架(Express/NestJS/Spring Boot/FastAPI/Django/Gin)的 路由识别、参数提取、校验规则提取、鉴权提取和响应结构提取模式。
内部步骤(Task 4),由 generator 编排器调用,不从用户直接触发。 从 PRD 分析和代码库映射结果生成精确的、编码 Agent 可执行的实现规格文档。
VictoriaTraces HTTP API reference for querying distributed traces via Jaeger-compatible API. This skill should be used when searching traces, listing services/operations, getting trace details by ID, querying service dependencies, constructing HTTP requests to VictoriaTraces, or working with Jaeger API endpoints for distributed tracing analysis.
MetricsQL query language reference for VictoriaMetrics-specific features like default_rollup, rollup, outlier detection, resource utilization functions (ru, ttf), time-series transformations, extended aggregations (topk_avg, outliers_iqr), and MetricsQL-only rollup functions. Use when writing queries with VictoriaMetrics extensions or when VictoriaMetrics query semantics (no extrapolation, optional lookbehind) are needed. For standard PromQL syntax and semantics, use the promql skill.
This skill should be used when the user asks to "explain how X works", "how does X work", "what is the implementation of X", "trace the code for X", "show me the call chain of X", "how is X implemented", "how does the X module work", "what's the design of X", "walk me through X", "explain the X feature", "what does X do", "how is X built", "解释一下 X 是怎么实现的", "帮我分析 X", "梳理一下 X 的设计", "X 的调用链是什么", "X 的代码流程是什么", or mentions understanding, analyzing, or exploring a feature, function, module, mechanism, or design in a codebase. Also triggers when the user asks about the architecture, flow, or internal workings of a component. Provides a structured analysis of how a feature is implemented. Do NOT use for debugging, bug fixing, or modifying code -- only for understanding and explaining.
This skill should be used when writing Postman test script assertions for HTTP responses, when the user asks to "assert status code", "validate response body", "check response headers", "extract variables from response", "write pm.test assertions", or mentions HTTP response verification in Postman test scripts.
This skill should be used when the user wants to "create a CNB issue", "list CNB issues", "close issue on cnb.cool", "update issue status", "add label to issue", "assign issue to someone", "comment on CNB issue", "查看 cnb issue", "创建问题单", "查询 CNB 问题", "更新 issue 状态", "给 issue 加标签", "分配 issue 处理人", "在 cnb.cool 上创建工单", or mentions managing issues on the CNB (cnb.cool) platform. Provides a CLI tool for full Issue lifecycle management. Do NOT use for GitHub or GitLab issues.
API 测试方法论知识库。提供系统化的测试用例设计方法:正向测试、反向参数校验、 安全测试和链路测试。每类测试包含用例命名规范、请求数据构造方法和预期结果断言。
内部步骤(Task 1),由 generator 编排器调用,不从用户直接触发。 分析用户原始输入,自适应选择工具加载 PRD 内容到工作目录。
内部步骤(Task 2),由 generator 编排器调用,不从用户直接触发。 对 PRD 内容进行 5-zone 结构化分析,提取功能、用户故事、业务实体、约束和验收条件。
This skill should be used when generating, validating, or editing Postman Collection v2.1.0 JSON files, when writing Postman test scripts (pre-request or test events), when structuring collection items, folders, variables, or auth configuration, or when the user asks to "create a postman collection", "generate postman tests", "validate collection format", or mentions "postman schema" or "v2.1.0".
This skill should be used when writing Postman test script assertions for VictoriaTraces call chain verification, when the user asks to "assert trace path", "verify call chain", "check span relationships", "validate distributed trace", "query VictoriaTraces from Postman", or mentions trace/span verification in API tests.
内部步骤(Task 3),由 generator 编排器调用,不从用户直接触发。 将 PRD 需求映射到代码库,识别技术栈、相关模块、编码约定和扩展点。
VictoriaMetrics HTTP API reference for querying metrics, exporting/importing data, TSDB stats, and administrative operations. This skill should be used when constructing HTTP requests to VictoriaMetrics, understanding query endpoints (/api/v1/query, /api/v1/query_range, /api/v1/export, /api/v1/import), response formats, checking cardinality, creating snapshots, or integrating with VictoriaMetrics API.
VictoriaLogs HTTP API reference for querying logs, hits stats, field discovery, live tailing, and log statistics. This skill should be used when constructing HTTP requests to VictoriaLogs, understanding query endpoints (/select/logsql/query, /select/logsql/tail, /select/logsql/hits, /select/logsql/field_names), response formats, or integrating with VictoriaLogs API for log search and analysis.
PromQL query language fundamentals for Prometheus and Prometheus-compatible systems. Use for understanding PromQL instant/range vectors, label matchers, aggregation operators, offset/@ modifiers, or when targeting non-VictoriaMetrics Prometheus systems. For VictoriaMetrics-specific features like default_rollup, rollup, or outlier detection, use the metricsql skill.
This skill should be used when writing Postman test script assertions for VictoriaLogs business log verification, when the user asks to "assert log events", "verify business logs", "check domain.action logs", "query VictoriaLogs from Postman", "validate log entries", or mentions business log verification in API tests.
LogsQL query language reference for VictoriaLogs. This skill should be used when writing LogsQL queries, searching logs in VictoriaLogs, filtering log streams, using pipe operators for log aggregation, building log analysis queries with stats/sort/fields pipes, or understanding LogsQL filter syntax including word search, regex, and field-level filters.
将产品需求文档(PRD)转化为精确的、编码 Agent 可直接执行的实现规格文档。 触发:用户要生成 spec / 规格文档 / 实现方案,或要把 PRD 转成可执行规格。 不触发:写 PRD、改 PRD、代码 review、或非文档生成任务。
内部步骤(Task 5),由 generator 编排器调用,不从用户直接触发。 对规格文档进行结构化质量审查,检查 PRD 覆盖度、任务可操作性和验收条件可测试性。
Postman Collection 构建知识库。提供 Postman v2.1.0 Collection JSON schema、 请求项结构、断言代码模板和环境变量模板。