How do I install the Bootstrap skill?

Run npx skillmds@latest add zedarvates/bootstrap in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Bootstrap safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Bootstrap?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Bootstrap free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Bootstrap?

zedarvates (@zedarvates) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Bootstrap by zedarvates — Agent Skill

Name: Bootstrap
Author: zedarvates

Bootstrap

Deploy Botte Secrète's token-saving stack into a target project — wire the botte-llm MCP server into .mcp.json, audit the project's agent directives, and write a .botte config + setup report. Use when the user wants to "install botte", "set up token savings on a project", reduce an existing project's token/cost usage, or onboard a repo to local-first routing. This is the capstone that makes the toolkit actually save money on real projects.

zedarvates 87717d4 5 files · 19.0 KB Updated Sep 17, 2026

File contents

bootstrap — deploy the toolkit into a project

The whole point of Botte Secrète is to make real projects cheaper to work on. This installs the stack into any project in one command.

Run it

python -m skills.bootstrap.cli /path/to/project
python -m skills.bootstrap.cli /path/to/project --create-agents-md   # scaffold AGENTS.md if missing
python -m skills.bootstrap.cli /path/to/project --scan-subnet --json

What it does (idempotent)

Discover local backends (llm_backends) so routing has somewhere to send work.
Wire MCP — merges the botte-llm server into the project's .mcp.json (non-destructive; keeps any existing servers). The project's agent then gains: auto_route, local_chat, fusion, find_skills, audit_local_usage, route_task, discover_backends, list_models.
Audit directives (directives_audit) — reports the CLAUDE.md/AGENTS.md health; with --create-agents-md, scaffolds a starter AGENTS.md when none exists.
Write .botte/config.json — chosen local model, detected cloud keys, token budget, routing mode (auto).
Write .botte/setup-report.json + print a summary with next steps.

Result

The project's agent now routes cheap work (classification, extraction, summaries, tool/skill search) to local models for 0 cloud tokens, escalates only the hard parts to the cloud (auto_router), and picks tools with find_skills instead of reading every skill description with the expensive model.

.mcp.json and .botte/ hold per-machine absolute paths, so they are gitignored in this repo — deploy on each machine.

Related: [[llm_mcp]] (the tools wired in), [[auto_router]], [[skill_finder]], [[directives_audit]], [[llm_backends]].

---name: bootstrapdescription: Deploy Botte Secrète's token-saving stack into a target project — wire the botte-llm MCP server into .mcp.json, audit the project's agent directives, and write a .botte config + setup report. Use when the user wants to "install botte", "set up token savings on a project", reduce an existing project's token/cost usage, or onboard a repo to local-first routing. This is the capstone that makes the toolkit actually save money on real projects.---# bootstrap — deploy the toolkit into a projectThe whole point of Botte Secrète is to make *real projects* cheaper to work on.This installs the stack into any project in one command.## Run it```bashpython -m skills.bootstrap.cli /path/to/projectpython -m skills.bootstrap.cli /path/to/project --create-agents-md # scaffold AGENTS.md if missingpython -m skills.bootstrap.cli /path/to/project --scan-subnet --json```## What it does (idempotent)1. **Discover local backends** (`llm_backends`) so routing has somewhere to send work.2. **Wire MCP** — merges the `botte-llm` server into the project's `.mcp.json` (non-destructive; keeps any existing servers). The project's agent then gains: `auto_route`, `local_chat`, `fusion`, `find_skills`, `audit_local_usage`, `route_task`, `discover_backends`, `list_models`.3. **Audit directives** (`directives_audit`) — reports the CLAUDE.md/AGENTS.md health; with `--create-agents-md`, scaffolds a starter `AGENTS.md` when none exists.4. **Write `.botte/config.json`** — chosen local model, detected cloud keys, token budget, routing mode (`auto`).5. **Write `.botte/setup-report.json`** + print a summary with next steps.## ResultThe project's agent now routes cheap work (classification, extraction, summaries,tool/skill search) to **local models for 0 cloud tokens**, escalates only the hardparts to the cloud (`auto_router`), and picks tools with `find_skills` instead ofreading every skill description with the expensive model.`.mcp.json` and `.botte/` hold per-machine absolute paths, so they are gitignoredin this repo — deploy on each machine.Related: [[llm_mcp]] (the tools wired in), [[auto_router]], [[skill_finder]],[[directives_audit]], [[llm_backends]].

Frequently asked questions