Menu
SkillsMP has collected 21 skills from hylarucoder/hai-stack. Open a skill to review its source and details.
Showing 21 of 21 collected skills.
Reviews a code change for correctness, security, maintainability, architecture, performance, and verification gaps, or performs a local code-smell review. Use for review this diff/PR、变更评审、代码审查、代码整洁度. Return evidence-backed findings and a scoped verdict;…
Audits an entity model field by field: what should exist, what exists now, where each value belongs (column, config, computed runtime value, or removal), and what must migrate. Use when comparing a PRD with a schema/model, designing entity fields, or deciding…
Challenges an overly conservative, incremental, or compatibility-led proposal with a bold target model, a concrete kill/merge/split list, materially different options, and a falsifiable first proof point. Use when the user asks to think bigger, open the…
Investigates architecture and software change complexity, from whole-repository runtime paths to a bounded module/design decision. Use for 系统太绕、复杂度审计、调用链/配置太散、架构审查、模块边界、拆分合并, or why a repo is hard to change. Select global investigation or bounded design mode;…
Produces a ready-to-run ast-grep command or reusable YAML lint/codemod rule, validated against positive and negative fixtures. Use when search or rewrite depends on syntax structure: finding all call shapes, avoiding regex matches in comments/strings,…
Audits documentation for internal contradictions and/or mismatches with code, config, schemas, and approved contracts. Use for 文档审计、前后冲突、术语漂移、README 是否过时、文档和代码一致吗, including check-and-fix requests. Select evidence sources from the request; return prioritized…
Diagnoses an unexplained software failure through reproduction, competing hypotheses, discriminating checks, and a causal explanation. Use for 排障、根因分析、偶发失败、环境差异、重复执行, or a bug whose cause is unclear. Return observed evidence, ruled-out hypotheses,…
Creates or rewrites a goal document for ambiguous, multi-phase, high-risk, or explicitly planned execution: verifiable outcome, boundary, current state, phases, proof-bearing todos, dry-run, and Go/No-Go. Use when the user asks for a goal, plan, phases,…
Decides whether a PRD is needed, drafts one from a product problem, repairs an existing PRD, or judges its split/merge boundary. Use for PRDs, product requirements, requirement specs, scope, user-visible behavior, and acceptance criteria（写 PRD、需求文档、PRD…
Rebuilds a drifted, repeatedly patched document from verified current conclusions. It inventories old blocks, preserves verified wording, removes disproved material, writes authoritative replacement coverage only from verified sources, and reports…
Diagnoses one React component or component directory across consumer API, data flow, testability, extensibility, performance, mental model, and boundaries/contracts, with code evidence and prioritized recommendations. Use when the user points to a component…
Reformats and restructures a README or similar Markdown project document without changing its factual content, voice, or scope, then summarizes formatting-only edits. Use when the user asks to beautify, tidy, organize, or professionally format a README（README…
Verify a defined software change against requirements and produce a proportional, evidence-backed acceptance report with specification precedence, scope boundaries, requirement-to-evidence traceability, executed test results, risk findings, a release verdict,…
Creates one decorative, single-frame visual card from supplied content as self-contained HTML and a shareable PNG. Use for information, quote, knowledge, social, summary, or cover cards, including “把要点做成一张图” and “做张卡片”. Use hai-visual-report for multi-section…
Pressure-tests an ambitious proposal and returns a landing judgment, one minimum viable proof, explicit cuts, success/failure signals, and a stop rule. Use when a direction has more vision than executable grounding or the user asks how to land it, reduce…
Evaluates whether an idea, feature, product, or project deserves attention and returns one verdict—Do, Validate first, Reframe, Defer, or Kill—plus the strongest objection and cheapest decision-changing test. Use when the user asks whether something is worth…
Recommends names for concepts, identifiers, modules, files, entities, events, or product terms, or audits an existing naming set for misleading vocabulary and viewpoint drift. Use for “what should I call this?”, rename requests, or naming…
Tests whether each requirement, step, role, field, state, module, layer, abstraction, or rule deserves independent existence, then classifies it as Keep, Merge, Defer, Delete, Replace, or Prove first. Use when the explicit decision is what can be removed or…
Diagnoses single-source-of-truth violations—duplicated literals or rules, shape drift, scattered defaults, redundant conversion chains, overloaded vocabulary, and behavior forks—with file:line evidence, false-positive adjudication, and a treatment recipe. Use…
Drives a real behavior change through red-green-refactor and records the failing RED evidence, minimal GREEN implementation, refactor decision, and verification commands. Use when the user explicitly asks for TDD/tests first, adds or fixes testable behavior,…
Turns an idea, PRD, plan, review, architecture topic, or other substantive source into a self-contained, multi-section HTML report with structure-appropriate diagrams, matrices, evidence, risks, and next moves while preserving the source’s meaning. Use when…