Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add toBzh30/Claude-Project-Bootstrap --skill split-claudemdgit clone --depth 1 https://github.com/toBzh30/Claude-Project-BootstrapWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/tobzh30/claude-project-bootstrap/split-claudemd)<a href="https://agentmods.dev/skills/tobzh30/claude-project-bootstrap/split-claudemd"><img src="https://agentmods.dev/badge/skills/tobzh30/claude-project-bootstrap/split-claudemd.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00060 | $0.02127 |
| Opus 5 | $0.00030 | $0.01064 |
| Sonnet 5 | $0.00012 | $0.00425 |
| Haiku 4.5 | $0.00006 | $0.00213 |
Grade A, and why
split-claudemd scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 6d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 175 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/split-claudemd — Refactor CLAUDE.md into hub-and-spokes
The goal: a short root CLAUDE.md that anyone can read in 30 seconds, with longer detail pushed to per-subdirectory CLAUDE.md files (Claude Code auto-loads these when the cwd is inside that subtree) and topic files under .claude/rules/.
Hard constraint — root CLAUDE.md only. Anthropic truncates the project root CLAUDE.md at 200 lines as of writing. Anything past the cap is silently dropped from the model's context — no warning to the user. Target ≤150 lines for the root to leave headroom. If you suspect the cap has changed, verify against current Claude Code docs before relying on the number.
Sub-CLAUDE.md files (frontend/CLAUDE.md, backend/CLAUDE.md, etc.) are not subject to the 200-line cap — they auto-load in full when Claude Code's cwd is inside that subtree. No hard cap, but keep them under ~250 lines as a soft target so they stay scannable and don't become new monoliths.
Topic files under .claude/rules/ are not auto-loaded — they're read on demand when something points to them. No cap, no soft target, but split if a single topic file turns into a reference manual.
This skill proposes a split and waits for confirmation before writing. The right split is project-specific — never auto-refactor.
Step 0 — Working tree check
This skill writes files (root CLAUDE.md, new spokes, topic files). Before any changes, give the user a clean baseline to git diff against if something goes wrong:
git status --short
If the output is non-empty, halt and report:
Uncommitted changes in
<N>files:<git status --short output, truncated to 10 lines if longer>This skill rewrites your root
CLAUDE.mdand may write new spoke files. If anything goes wrong you'll want a clean baseline. Commit/stash first, or continue anyway?
Wait for explicit "continue" before proceeding. Don't proceed silently.
If git status --short errors (not a git repo), warn the user and ask whether to proceed without a revert baseline. Recommend git init first.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 6d ago First seen · 175 lines · 60 tokens per session scan A 14d37689ba5c
split-claudemd is a skill published in the GitHub repository toBzh30/Claude-Project-Bootstrap (3 stars, last pushed 1mo ago), licensed MIT. It adds 60 tokens to every session and 2,127 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
project-standard-spec
Spec maintainer for Big Emotion Project Standard. Reads Confluence (Requirements / Decisions / Architecture) as the source of truth, helps you investigate a problem, structures the reflection, and produces drafts — Pending REQ/DEC/ARCH sections on Confluence + matching Jira tickets in the Jira project configured in…
fix-ci-train
Batch-fix the CI on every open pull request whose required checks are red — CI only, no review, no merge. From a GitHub PRs URL (or the current repo), enumerate every open PR with failing required checks, fetch each PR's failing-check logs, consolidate a per-PR CI-fix checklist, dispatch a team of parallel fixer…
fix-train
Batch-apply the requested changes on every open pull request in a GitHub list — fix only, no merge. From a GitHub PRs URL (or the current repo), enumerate every open PR, collect each PR's requested changes (formal CHANGESREQUESTED reviews, inline review comments, and "Issues requiring changes" comments /…
issue-train
Batch-process every open GitHub issue in a list — triage, detect existing PR linkage, detect silent fixes in already-merged PRs, close issues that are no longer pertinent, then plan + build + open one PR per remaining actionable issue with Closes.
blueprint
Extract a complete specs/ directory from any frontend codebase. Covers all 12 axes needed to rebuild or migrate a frontend app. Produces 80% quality on first pass, refined to 95% through guided iteration.
project-standard-audit
Production-readiness audit for Big Emotion Project Standard (this repo). Read-only multi-domain scored assessment that answers four questions — is the plugin ready to install on real repos, is the template surface healthy, what is the security posture, and is the score close to 8–9/10. Use when the user asks "is it…