Recover Semantic Model
Recover a program's minimal semantic model (the state that changes what users see, how its transitions repeat and reorder, what each effect requires). Use when asked to recover or audit that model, to check what is idempotent or order-dependent, or to split a refactor into provably equivalent and "trust me" changes.
Install
npx skills add kr4chinin/skills --skill recover-semantic-modelgit clone --depth 1 https://github.com/kr4chinin/skills.git /tmp/skills
mkdir -p ~/.claude/skills
cp -r /tmp/skills/skills/recover-semantic-model ~/.claude/skills/recover-semantic-modelPersonal skills live in ~/.claude/skills. Use .claude/skills in a project folder to share a skill with your team.
From the SKILL.md
# Recover the semantic model Recover the program's minimal semantic model: the smallest description of its state, transitions and dependencies that predicts what its users observe. Every finding points at a row of the model; one that names no row is ordinary code review. Stay read-only unless asked: no edits, tests or tickets. By default build the three tables and run the four passes; split refactors when the request mentions them or names a diff. Asked for one part, do that part and the rows it needs. Read the [report layout](references/report.md) before the first table: its worked example shows how full each cell should be. ## Scope and screen Use the scope the user gave; with several programs and no hint, ask which. Report what you modelled, what you skipped and the revision you read. The screen is where a person, or a system acting for them, observes the program: what a UI renders; a service's responses, messages, records other programs read and money it moves; a library's returns, throws and callbacks. Number rows S1, T1, E1, one row per stored copy of a fact, not per variable. On a large scope, model one lifecycle or component at a time (a job, a document, a family of consumers), main flows first, and list the rest as not modelled. ## 1. State: every value that changes what the user sees Stored fields, in-memory and client state, values held by another system (a queue, a cache, a workflow engine), configuration that selects behaviour.
Details
- Skill name
recover-semantic-model- Repository
- kr4chinin/skills
- Path
skills/recover-semantic-model/SKILL.md- License
- MIT
- Repo updated
- 16 November 2026
FAQ
What does the Recover Semantic Model skill do?
Recover a program's minimal semantic model (the state that changes what users see, how its transitions repeat and reorder, what each effect requires). Use when asked to recover or audit that model, to check what is idempotent or order-dependent, or to split a refactor into provably equivalent and "trust me" changes.
How do I install the Recover Semantic Model skill?
skills CLI (Claude Code, Cursor, Codex and more): npx skills add kr4chinin/skills --skill recover-semantic-model. Skills go in ~/.claude/skills (personal) or .claude/skills (project), and Claude loads them automatically when a task matches the description.
Is Recover Semantic Model popular?
It lives in kr4chinin/skills, which has 9 GitHub stars and 0 forks.
Related skills
Convert State Machines to Actors
9kr4chinin/skills
Convert a database-backed implicit state machine to an actor, the one owner every writer of the persisted state goes through. Use when asked to put writers behind one owner, or to fix a race or lost update across several writers, such as a find-implicit-state-machines finding.
Find Implicit State Machines
9kr4chinin/skills
Audit a codebase for implicit state machines (state spread across flags, status fields, timers and callbacks). Use when asked to find them or to trace impossible states, stuck loading, missing resets or async races. Catalog machines with no defect only when asked for an inventory.
Prune Comments
9kr4chinin/skills
Audit code comments and plan which to delete, keep, rewrite or replace with a refactor. Use when asked to prune a codebase's comments, or to sort workaround comments into root-cause fixes and safe deletions.
Storyblok Use MCP
9storyblok/agent-plugin
Use when interacting with Storyblok through the MCP server — the tool model, operation names and payload shapes for stories, assets, components, datasources and every other resource. Load it before calling `mcp__storyblok__` tools, alongside the skill that owns the task.
AI Paper Writing
9HOWILLMAKEIT/skills
辅助 AI 论文写作,分两类任务。一是论文主图、方法图、机制图和图形摘要:搜集顶会同类型参考图,写网页生图提示词,用户手动生图后重建为原生可编辑 PPT。二是摘要、Introduction、实验等文字段落:先搜集相关领域顶会的对应段落到文件夹(md),暂停;用户需要时再阅读并总结写法与共同逻辑,暂停;再按总结的逻辑顺序,结合用户提供的代码和论文草稿完成写作。文字写作的三步各自暂停,等用户确认后才进入下一步。实测数据图应从真实数据绘制。
Agent Builder
5NikolasMarkou/agent-builder
Build, review, troubleshoot, optimize, and extend AI agents — single-agent, multi-agent, RAG, and production systems. Covers framework selection, pattern selection (topology, behavioral, data flow), agent orchestration, state design, and tool design. Defaults to Python with LangChain/LangGraph; also supports CrewAI, Strands, OpenAI Agents SDK, Google ADK, and Mastra. Use whenever the user wants to build, design, scaffold, review, debug, optimize, extend, or migrate an agent or agentic system.
Browse the most popular Claude skills, all ai, llms & agents skills, or write your own with the skills generator.