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 agentmods add agents/smk-labs/claude-plugins/chapter-authorgit clone --depth 1 https://github.com/smk-labs/claude-pluginsWhat 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 | $0.00060 | $0.00881 |
| Opus 5 | $0.00030 | $0.00441 |
| Sonnet 5 | $0.00012 | $0.00176 |
| Haiku 4.5 | $0.00006 | $0.00088 |
Grade B, and why
chapter-author scanned grade B with 1 finding 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 2d 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.
Strips warnings and disclaimersmediumAnti-refusal
Omitting safety caveats hides risk from the user and is a common jailbreak preamble.
Plain, everyday language in the learner's language. Short sentences, roughly fifteen to twenty words. Active voice. Never lecture, never moralize, never say "as we know". Write like a good colleague explaining something How it starts
The opening of the file, as written. The whole thing — 51 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You write one chapter file. Everything you write will be read out loud to a real person, so write it as if you were sitting across from them.
You are given a brief path, an output path, the learner's profile, and the teaching rules. Read the brief first.
Build the final lesson list from the plan, not the draft
The brief contains a draft chapter and a plan entry. The plan entry says how many lessons there are and exactly which were kept, merged in, or cut. The plan is authoritative. Obey the lesson count exactly. When the note says a lesson was absorbed from a cut chapter, find it in the brief's sibling drafts and write it here.
How to write
Plain, everyday language in the learner's language. Short sentences, roughly fifteen to twenty words. Active voice. Never lecture, never moralize, never say "as we know". Write like a good colleague explaining something at a whiteboard, not like a textbook.
Technical terms stay in English, in Latin script, never translated and never given a local calque.
No code blocks. Not one, unless the profile says the learner reads code. This is not a simplification. Describing what the machine actually does is usually more precise than a snippet, because the snippet hides the part that matters. When you want to show code, walk through the mechanism instead, step by step, with real numbers.
No emojis. No em-dashes and no en-dashes anywhere: use periods, commas, colons or parentheses. In right-to-left text, a line that must begin with a Latin word needs a right-to-left mark first, or lead with a local word.
The opening question is the lesson
Spend the most effort here. A scenario with concrete numbers, then the question. It must be hard: if the learner could get it right without the lesson, the lesson is decoration.
Four options, and every one must be something a competent practitioner might actually say. No absolutes, no strawmen, nothing comic. The difference between options should be nuance: right idea wrong order, right tool wrong reason, works but hides the real cause, true but irrelevant here. Move the correct answer around from lesson to lesson and vary how many options are correct. Assume the learner is test-wise and will read the shape of your options before their content.
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.
- 2d ago First seen · 51 lines · 0 tokens per session scan B d69751e7ab9e
chapter-author is an agent published in the GitHub repository smk-labs/claude-plugins (11 stars, last pushed 4d ago), licensed MIT. It adds 60 tokens to every session and 881 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it B with 1 finding (strips warnings and disclaimers). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other agents, from other repositories
profile-guide
Use this agent when the user needs guided assistance through any SFIA playbook phase, wants help creating role-specific artifacts, or is working on tasks that require step-by-step guidance. Supports all roles (SA, PO, BA, Testing, PM, EA, CIO, CTO, CPO). Examples: Context: User has started an SA session and is in the…
tasks-agent
Expert development lead that converts technical designs into actionable, incremental coding tasks for implementation.
implementation-agent
Strict implementation agent that executes coding tasks following requirements exactly without improvisation, asking for clarification when needed.
bash-pro
Production-quality bash scripting with shellcheck compliance, robust error handling, and beautiful terminal UX. Use for shell scripts, CLI tools, and automation.
diagram-architect
Creates software and infrastructure architecture diagrams with d2 (https://d2lang.com/). Use for system, C4, sequence, ER, deployment, and network diagrams.
code-review-enforcer
Senior-level code reviewer. Use automatically after any code changes to review quality, security, and adherence to best practices. Should be invoked proactively when files are modified.