feature

A complete writing workflow for The Claude Code Bible, a Markdown book about Claude Code. It covers checking claims, dividing chapters among agents, writing, fixing links and tables, and merging the finished work.

In plain words
What is it for?
Use it to research and draft chapters, update the table of contents and cross-links, and commit and merge book changes by file path.
Why use it?
It coordinates the many steps needed to publish readable chapters when several agents edit one repository. It also defines when work is actually finished and when a problem requires human input.

Command for Claude Code

Install

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.

agentmods
npx agentmods add commands/sebyx07/claude-code-bible/feature
Clone the repo
git clone --depth 1 https://github.com/sebyx07/claude-code-bible

Made for: Claude Code.

Per session 59 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 3,355 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00059 $0.03355
Opus 5 $0.00030 $0.01677
Sonnet 5 $0.00012 $0.00671
Haiku 4.5 $0.00006 $0.00335

Measured 3d ago against content hash 234521f50786, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

feature 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 3d 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.

.claude/commands/feature.md · 101 lines

How it starts

The opening of the file, as written. The whole thing — 101 lines — stays where its author put it; the contents beside it link to each section on GitHub.

/feature

You are an author on The Claude Code Bible — a prose book shipped as Markdown. No code, no build, no tests, no CI. Edits ship by committing .md files.

Done means merged and readable on GitHub — nothing less counts. There is no build to go green and no release process: readers consume main directly, so the arc is understand → verify → slice → draft → check every link and table → merged to main. A drafted chapter is not done; an open PR is not done; a merged chapter with one broken relative link is a chapter readers bounce out of. When you report, say what you actually rendered and clicked versus what you assume works.

Request

$ARGUMENTS

The prompt is the context — read the intent. "Just write it" / "do full work" → research, draft, fix the TOC and cross-links, commit and merge on your own judgement; put the editorial calls in the PR body instead of asking. A tentative ask ("should we cover X?") → settle the scope first, then write. Always stop for a true blocker: a claim about Claude Code you cannot verify, a renumbering that would rewrite every inter-chapter link, or a rewrite of a chapter someone else is visibly mid-edit on.

Pick the PR mode before you brief anyone. Chapter-per-PR (default) or one PR for a coherent sweep (a rules change that ripples through every chapter's examples). Path-disjointness governs the drafting either way — it is how parallel agents avoid clobbering each other.

Cap a PR at ~5 chapters. The ~110–120 file cap scales down hard here: a book PR is read, not diffed. Past a handful of chapters a reviewer skims, one contested paragraph holds four finished chapters hostage, and a later git blame lands on one enormous "docs: rewrite". Split it even if the user asked for one PR, and say why. Renumbering is the exception that must never be split — renaming files and rewriting every cross-link is one atomic commit or a broken book.

Work as a hive mind, in one checkout

Hiving is a judgement call, not a ritual. Only two things justify it: searching (sweeping all 13 chapters for a claim, a stale version string, a broken link — you want conclusions, not file dumps) and scale (several independent chapters to draft or revise). One chapter, one typo, one retitle: do it yourself. Briefing an agent to fix a sentence costs more than the sentence.

Read the full file on GitHub · 101 lines

Changes

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.

  1. 3d ago First seen · 101 lines · 59 tokens per session scan A 234521f50786

Subscribe to this mod's changes

feature is a command published in the GitHub repository sebyx07/claude-code-bible (5 stars, last pushed 1mo ago), licensed MIT. It adds 59 tokens to every session and 3,355 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.