Instructions for subagents, separate helper agents started through the Task tool, with four work lanes and read-only defaults.

Use it when coordinating subagents through the Task tool and dividing work across four lanes.

It defines how these helper agents are expected to operate and limits their default access to changes.

README — Agent for Claude Code by shawnclybor/clybor-claude-tooling

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.

Clone the repo

git clone --depth 1 https://github.com/shawnclybor/clybor-claude-tooling

Made for: Claude Code.

Wrote 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.

README.md

[![agentmods](https://agentmods.dev/badge/agents/shawnclybor/clybor-claude-tooling/readme/github.svg)](https://agentmods.dev/agents/shawnclybor/clybor-claude-tooling/readme)

Your own site

<a href="https://agentmods.dev/agents/shawnclybor/clybor-claude-tooling/readme"><img src="https://agentmods.dev/badge/agents/shawnclybor/clybor-claude-tooling/readme/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

Your own site · 80×15

<a href="https://agentmods.dev/agents/shawnclybor/clybor-claude-tooling/readme"><img src="https://agentmods.dev/badge/agents/shawnclybor/clybor-claude-tooling/readme.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 0 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.

When invoked 2,045 The whole file, excluding the scripts and references it only reads on demand.

Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.

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.

Model	Per session	Once invoked
Fable 5.1	$0.00000	$0.02045
Opus 5.5	$0.00000	$0.00818
Sonnet 5.5	$0.00000	$0.00409
Haiku 4.5	$0.00000	$0.00204

Measured 18d ago against content hash b9cd9f55d250, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

README 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 18d 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/agents/README.md · 83 lines

How it starts

The opening of the file, as written. The whole thing — 83 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Agents

Nineteen subagents in four lanes, spawned with the Agent tool. Every file carries name, description and model in its frontmatter; none restricts tools, so each agent has whatever the session has. Read-only by convention: an agent reports, and the skill that spawned it decides what lands.

Model tiers

The frontmatter uses the aliases fable, opus, sonnet, haiku, which resolve to the newest model of each tier: Fable 5.1, Opus 5, Sonnet 5, Haiku 4.5. The rule for picking one:

Tier	Use it for	Why
Opus 5	A verdict that is wrong expensively: is this claim true, is this surface secure	Judgment work. Two agents run on it: `adversarial-reviewer` and `security-auditor`
Fable 5.1	No agent by default	The most capable model available, and priced for it. Reviews run three agents a round and rounds repeat, so a default here compounds. When one verdict warrants it, pass `model: "fable"` on that Agent call and leave the file alone
Sonnet 5	Structured work with a clear procedure: a lens with one question, a code review against a checklist, a reproduction, a synthesis, a citation check	Fast, cheap enough to run three in parallel, and the procedure carries the quality
Haiku 4.5	One question, one cited answer	A lookup does not need judgment

The frontmatter is the default for every project that inits from here, so it stays at the tier a routine round can afford.

An agent that no skill, command or hook names does not get used. Sessions reach for the tools a skill puts in front of them, not for a roster. So the useful column below is Called by, measured on 2026-09-04 by grepping every skill, command and hook in this repo for the agent's name. Four agents are called by nothing; they are listed last, on purpose, and are the first candidates to cut.

Quality (`quality/`)

Three non-overlapping lenses. Each gets one question, and a finding outside its column is dropped rather than reported; three agents reporting one defect is three counts of one problem. The boundaries are written down in .claude/skills/adversarial-review/rubric.md.

Read the full file on GitHub · 83 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.

18d ago Changed b9cd9f55d250
2mo ago Changed · +21 lines c6bb430d625a
2mo ago First seen · 62 lines · 0 tokens per session scan A 5cd598606b47

Subscribe to this mod's changes

README is an agent published in the GitHub repository shawnclybor/clybor-claude-tooling (6 stars, last pushed yesterday), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 2,045 tokens. 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.

Building agents adversarial-review agent-orchestration ai-agents ai-development anthropic boilerplate

README next to 6 agents from other repositories
Mod	Type	Stars	Last push	Licence	Scan	Tokens	Works with
READMEshawnclybor/clybor-claude-tooling	Agent	6	yesterday	MIT	A	0	Claude Code
plan-checkerSienkLogic/plan-build-run	Agent	18	7mo ago	MIT	A	26	Claude Code
debuggerSienkLogic/plan-build-run	Agent	18	7mo ago	MIT	A	24	Claude Code
auditSienkLogic/plan-build-run	Agent	18	7mo ago	MIT	A	44	Claude Code
integration-checkerSienkLogic/plan-build-run	Agent	18	7mo ago	MIT	A	31	Claude Code
nyquist-auditorSienkLogic/plan-build-run	Agent	18	7mo ago	MIT	A	22	Claude Code
generalSienkLogic/plan-build-run	Agent	18	7mo ago	MIT	A	22	Claude Code