Tlaplus From Source

Generate a high-level TLA+ model from source code (C, C++, Rust, etc.). Analyzes code to understand its purpose, creates abstractions, writes TLA+ specification, and proposes invariants and properties. Use when the user wants to model source code in TLA+, create a formal specification from implementation, or verify concurrent/distributed algorithms.

49 stars 1 forksSoftware Development

Install

skills CLI (Claude Code, Cursor, Codex and more)
npx skills add tlaplus/AgentSkills --skill tlaplus-from-source
Manual install
git clone --depth 1 https://github.com/tlaplus/AgentSkills.git /tmp/AgentSkills
mkdir -p ~/.claude/skills
cp -r /tmp/AgentSkills/skills/tlaplus-from-source ~/.claude/skills/tlaplus-from-source

Personal skills live in ~/.claude/skills. Use .claude/skills in a project folder to share a skill with your team.

From the SKILL.md

# Generate High-Level TLA+ Model from Source Code

Create a TLA+ specification from source code by understanding the code's intent, abstracting implementation details, and focusing on the essential concurrent/distributed behavior.

## Philosophy

TLA+ models should capture **what** the system does, not **how** it does it:
- Focus on state transitions and their effects
- Abstract away implementation details (memory management, error handling boilerplate)
- Identify the core concurrent/distributed behavior worth verifying
- Keep the model small enough to be tractable for model checking

## Workflow Overview
Read the full skill on GitHub

Details

Skill name
tlaplus-from-source
Path
skills/tlaplus-from-source/SKILL.md
License
MIT
Repo updated
4 February 2026

FAQ

What does the Tlaplus From Source skill do?

Generate a high-level TLA+ model from source code (C, C++, Rust, etc.). Analyzes code to understand its purpose, creates abstractions, writes TLA+ specification, and proposes invariants and properties. Use when the user wants to model source code in TLA+, create a formal specification from implementation, or verify concurrent/distributed algorithms.

How do I install the Tlaplus From Source skill?

skills CLI (Claude Code, Cursor, Codex and more): npx skills add tlaplus/AgentSkills --skill tlaplus-from-source. Skills go in ~/.claude/skills (personal) or .claude/skills (project), and Claude loads them automatically when a task matches the description.

Is Tlaplus From Source popular?

It lives in tlaplus/AgentSkills, which has 49 GitHub stars and 1 forks.

Related skills

Tlaplus Add Variable

49

tlaplus/AgentSkills

Add a new variable to an existing TLA+ specification without changing its semantics. Ensures the variable is declared, initialized, and added to all UNCHANGED statements. Use when the user asks to add, introduce, or declare a new variable in a TLA+ spec, or mentions UNCHANGED statements.

Other

Tlaplus Split Action

49

tlaplus/AgentSkills

Split a TLA+ action into two sequential actions by introducing a new program counter (pc) state. Handles pc variable updates, UNCHANGED statements, TypeOk predicates, and follows naming conventions with renumbering. Use when the user asks to split, divide, or break an action into two parts, or wants to add an intermediate step to an action sequence.

Software Development

Programmed Illusion

113

Hmbown/Wizards-of-the-Ghosts

In D&D, Programmed Illusion creates an illusion that lies dormant until a specific trigger condition is met, then plays out its scene automatically. The real-world version is the conditional demo, automated preview, or triggered walkthrough: onboarding flows that appear when a user first visits, contextual tooltips that activate on hover, preview environments that spin up on pull request, automated scenario presentations that play when specific data conditions are detected. Programmed Illusion is the illusion that does not need a wizard present to cast it.

Software Development11

Game Code Review

57

AlterLab-IEU/AlterLab_GameForge

Reviews game code and AI-agent changes as a forked reviewer and writes a severity-ranked report: frame-rate independence, hot-path allocations, state and resource lifecycle, hallucinated engine APIs and leaked secrets, for Godot, Unity, Unreal and web. Use for "review my code", "review this PR", "check what the agent wrote". Not for design documents (use game-design-review).

Software DevelopmentIn 1 list

Mending

113

Hmbown/Wizards-of-the-Ghosts

Mending is surgical repair of a known-working artifact that has a specific, identifiable break. It is NOT: - Triage/diagnosis: If the problem is "API is throwing 500s, figure out why" — that's investigation, not mending. Mending requires the break to already be identified. - Formatting/cleanup: If the data is correct but looks messy (sort keys, normalize indentation) — that's formatting, not repair. - State reset: If a flag is stuck or an alert needs clearing — that's an operational toggle, not a file repair. - Refactoring/redesign: If the system is "fundamentally flawed" and needs redesign —

Software Development11

Game Godot Specialist

57

AlterLab-IEU/AlterLab_GameForge

Writes and reviews Godot 4 code (4.7 stable, 4.8 beta) in GDScript and C#: scenes and nodes, signals, Resources, autoloads, Jolt physics, shaders, profiling, export, GUT/gdUnit4 tests and headless CLI checks. Use when the project has project.godot, .gd or .tscn files, or the user mentions Godot or GDScript. Not for engine-agnostic architecture (use game-technical-director).

Software DevelopmentIn 1 list

Browse the most popular Claude skills, all software development skills, or write your own with the skills generator.