Claude Code Tresor: 141 Agents and 19 Slash Commands for Claude Code
A world-class collection of Claude Code utilities: autonomous skills, expert agents, slash commands, and prompts that supercharge your development workflow
At a glance
- What is it?
- Claude Code Tresor is a collection of 141 agents, 19 slash commands, skills, and workflow templates for Claude Code. It installs with a single shell command and is structured as a reference implementation for production-grade Claude Code workflows.
- Who is it for?
- Claude Code Tresor is a practical starting point for developers who want to move beyond ad-hoc Claude Code usage and into a structured, multi-agent workflow. The 19 slash commands and 141 agents cover most of the software development lifecycle, and the Tresor Workflow Framework provides a meta-prompting layer for complex orchestration.
- Can I use it commercially?
- Yes. MIT is a permissive licence: you can use, modify and sell software built on it, as long as you keep its copyright and licence notices.
- Is it still maintained?
- Yes. The repository last received commits 88 days ago.
- What is it written in?
- Mainly Shell, according to GitHub's language statistics.
Answers come from the project's GitHub data, last synced on September 29, 2026, and from our analysis. They are not legal advice.
Editorial analysis
What Claude Code Tresor Provides
Claude Code Tresor is a curated collection of configuration files for Claude Code, organized into agents, slash commands, skills, prompts, and workflow templates. The README describes the audience as solo developers who want tested tools for the full development lifecycle, teams routing work across multiple agents, and platform engineers building AI-assisted tooling who need a reference implementation.
The repository was created in September 2025 and reached version 2.7.0 in December 2025, at which point it included 141 agents in a `subagents/` directory organized into 10 color-coded team categories, plus 19 slash commands. The collection was available through both GitHub and Smithery. The last push to the repository was on 2026-07-03, about 87 days before the date of this review.
The 19 Slash Commands Across Four Categories
The 19 slash commands divide into four groups: Development, Tresor Workflow, Orchestration, and one implied general group. Four development commands cover common tasks: `/scaffold` for generating project structures and components, `/review` for code review with security and performance checks, `/test-gen` for generating test suites to a specified coverage target, and `/docs-gen` for generating documentation in formats including OpenAPI.
Five Tresor Workflow commands manage the meta-prompting layer: `/prompt-create` generates an optimized prompt for a complex task, `/prompt-run` executes a generated prompt in sub-agents with optional parallelism, `/todo-add` captures ideas without interrupting the current conversation, `/todo-check` reviews captured todos, and `/handoff-create` generates a context handoff document for session transitions.
Ten Orchestration commands added in v2.7.0 handle production-grade concerns. Security commands include `/audit` (2 to 4 hours, covering OWASP Top 10, pentesting, and infrastructure review), `/vulnerability-scan` (30 to 60 minutes, with CVE correlation and auto-fix), and `/compliance-check` (1 to 2 hours, covering GDPR, SOC2, HIPAA, PCI-DSS, ISO 27001, and CCPA). Performance commands include `/profile` (bottleneck analysis and Core Web Vitals) and `/benchmark` (load testing and capacity planning). Operations commands include `/deploy-validate`, `/health-check`, and `/incident-response`. Quality commands include `/code-health` and `/debt-analysis`.
Installing Claude Code Tresor
The README documents a one-command install approach. The exact install command is referenced in the README as a script in the repository. The general approach for collecting the agents and commands involves cloning the repository and running the install script, which copies the relevant files into the Claude Code configuration directory. The `GETTING-STARTED.md` file documents the full installation sequence.
Alternatively, the collection is available on Smithery for discovery and installation. The README points to three related repositories in the same ecosystem: the Claude Code Skill Factory for building custom skills, the Claude Skills Library for pre-built professional domain packages, and a comprehensive augmentation guide published as a GitHub Gist.
The `scripts/` directory contains verification scripts. The repository includes a `SCRIPTS-VERIFICATION.md` and a `VALIDATION-REPORT-CONTENT.md` (referenced in the README badge showing a quality rating of 9.7 out of 10).
The 141-Agent Structure and Team Categories
Version 2.5.0 reorganized all agents into a `subagents/` directory with 10 color-coded team categories. The 141 total agents consist of 8 core agents and 133 subagents. The v2.5.0 release also renamed three core agents from their earlier names: the architect became `@systems-architect`, the code reviewer became `@config-safety-reviewer`, and the debugger became `@root-cause-analyzer`. These renames were breaking changes in v2.5.0.
The orchestration system in v2.7.0 builds on this structure: the 10 new orchestration commands auto-select from the 141 agents based on the detected technology stack. Multi-phase orchestration runs 3 to 4 phases with parallel and sequential execution as appropriate. Dependency verification runs before parallel agent execution to ensure that agents writing to the same files do not conflict.
The README also documents a blog post on Medium titled "141 Claude Code Agents: The Setup That Actually Works" for a deeper explanation of the architecture decisions behind organizing this many agents.
The Tresor Workflow Framework
The Tresor Workflow Framework, added in v2.7.0, is a meta-prompting layer on top of the existing commands and agents. It provides todo management across a session (`/todo-add`, `/todo-check`), generated and stored prompts that can be rerun with `/prompt-run`, and a context handoff document via `/handoff-create` that preserves session state when transitioning to a new Claude Code context.
The `WORKFLOW-GUIDE.md` file in the repository documents how these pieces work together. The handoff document approach addresses the problem of context loss when a Claude Code session runs out of token budget: rather than starting from scratch, the handoff document contains the structured state of the previous session that the new session can load.
A comparable alternative is the Roo Code custom modes system, which allows defining agent personas with specific tool access and instruction sets. Claude Code Tresor differs by providing a larger ready-made collection with multi-phase orchestration rather than a framework for building custom modes from scratch.
Limitations and Version History
The collection's breadth is also its main limitation. A 141-agent system installed into Claude Code's configuration directory creates a large surface area. The orchestration commands document estimated durations ranging from 5 minutes to 4 hours, which means some commands represent substantial compute spend if triggered incorrectly. The README's guidance to review AI-generated code before submitting pull requests applies equally to AI-generated agent outputs.
The version history shows rapid iteration: v2.5.0 introduced breaking renames in November 2025, v2.6.0 improved design quality in the same month, and v2.7.0 added 10 orchestration commands plus the workflow framework in December 2025. The last push was in July 2026, suggesting the rapid initial phase has slowed. Version 2.7.0 added 12,682 lines of orchestration code according to the README.
The MIT license permits commercial use and modification. The README notes that the project is produced by a named author (Alireza Rezvani) and that all code is tested and used in real-world projects, though independent verification of that claim is not available through the repository.
Editorial conclusion
Claude Code Tresor is a practical starting point for developers who want to move beyond ad-hoc Claude Code usage and into a structured, multi-agent workflow. The 19 slash commands and 141 agents cover most of the software development lifecycle, and the Tresor Workflow Framework provides a meta-prompting layer for complex orchestration. The right time to skip it is when you only need one or two capabilities: adding 141 agents to a project for a single use case adds overhead without proportional benefit. Verify before installing that your use cases overlap with the commands and agent categories it provides. The last push was on 2026-07-03 and the project carries an MIT license.
Frequently asked questions
What is Claude Code Tresor and how is it different from plain Claude Code?
Claude Code Tresor is a collection of pre-built agents, slash commands, skills, and workflow templates that extend Claude Code. Plain Claude Code has no pre-configured subagents or slash commands; Tresor adds 141 agents and 19 commands on top of it.
Are the orchestration commands in Claude Code Tresor safe to run on a production codebase?
The README documents risk grades per ticket generated by the orchestration commands: low-risk quick wins, observe-with-recheck items, and high-risk technical changes that require backup and rollback discipline. The README also states that all code must be thoroughly reviewed before submitting.
Which version introduced the subagents directory and the 10 team categories?
Version 2.5.0 introduced the subagents/ directory with 10 color-coded team categories. That release also renamed three core agents, which were breaking changes documented in the MIGRATION-GUIDE.md.
Official sources
Add this badge to your README
If you maintain this project, the badge below links readers to this analysis and shows its maintenance status from the daily GitHub snapshot. Paste the markdown into your README; add ?metric=license or ?metric=stars to the image URL for a different field.
[](https://hysenlabs.com/projects/alirezarezvani-claude-code-tresor)