noisy-tools-in-subagent

A plugin that sends output-heavy commands, such as builds, tests, linters, and code analysis, to a secondary AI agent. A linter is a tool that checks code for likely mistakes or style problems.

In plain words
What is it for?
Use it when running builds, tests, linters, or static analysis that produce large amounts of output. Specific agent types can be exempted from delegation.
Why use it?
It keeps long command output out of the main agent's conversation context, leaving more room for the actual task and decisions.

Plugin 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.

Claude Code
/plugin marketplace add fprochazka/claude-code-plugins
agentmods
npx agentmods add plugins/fprochazka/claude-code-plugins/noisy-tools-in-subagent
Clone the repo
git clone --depth 1 https://github.com/fprochazka/claude-code-plugins

Made for: Claude Code.

Per session not measured What this adds to a session before it is invoked.
When invoked not measured Not applicable: nothing here is loaded into a session.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Security

Grade A, and why

noisy-tools-in-subagent 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 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.

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.

plugins/noisy-tools-in-subagent/.claude-plugin/plugin.json · 27 lines

What it actually says

{
  "$schema": "https://anthropic.com/claude-code/plugin.schema.json",
  "name": "noisy-tools-in-subagent",
  "version": "0.6.0",
  "description": "Forces noisy commands (builds, tests, linters, static analysis) to run inside a Sonnet subagent instead of the main context, preserving main-agent tokens.",
  "author": {
    "name": "Filip Procházka",
    "url": "https://github.com/fprochazka"
  },
  "homepage": "https://github.com/fprochazka/claude-code-plugins/tree/main/plugins/noisy-tools-in-subagent",
  "license": "MIT",
  "userConfig": {
    "exempt_agent_types": {
      "type": "string",
      "multiple": true,
      "title": "Exempt agent types",
      "description": "Agent types allowed to run noisy commands directly. Use for agents that cannot delegate, because their definition omits the Agent tool or they run at the subagent nesting limit. Each entry is a regex that must match the whole agent type, for example my-plugin:leaf-worker."
    }
  },
  "dependencies": [
    {
      "name": "bash-classify-hook",
      "marketplace": "fprochazka-bash-classify"
    }
  ]
}
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. 2d ago First seen · 27 lines scan A 0dc88d46904c

Subscribe to this mod's changes

noisy-tools-in-subagent is a plugin published in the GitHub repository fprochazka/claude-code-plugins (11 stars, last pushed 4d ago), licensed MIT. Its token cost is not measured: this kind of file is read by the harness, not the model. 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 2026-08-30.

Related

Other plugins, from other repositories

skarn-guard

Real-time guard on five Claude Code pre-execution hook events (tool calls, permission decisions, submitted and expanded prompts, and subagent creation) plus a response-end advisory that flags an action leaking a secret, installing a malicious package, or exfiltrating data, plus the skarn-audit skill (scan sessions and.

skarn-security/agent-guard · not measured

goalkeeper

Durable goal-following for Claude Code: contracts with definition-of-done, subagent judge gates, and linear chains.

bonfire-systems/goalkeeper · not measured

forge

A generator: it interviews an organization and emits a standalone, org-owned Claude Code plugin (the org's agent harness) aligned to how that org actually works. Karpathy-schema wiki + role/subagent pattern + skill verbs + tool adapters + the operating model. Re-emit to upgrade; never fork.

ktoulgaridis/forge · not measured

coaltipple

CoalTipple — model/effort router for Claude Code. Delegates a task you CAN do but is large+cheap DOWN to a lower tier to save tokens; escalates a task beyond the current tier UP for quality — up to Fable 5, the top rung (a consent-gated, real-money spawn). Two-knob routing (tier × effort, effort-before-tier), an alias.

TheColliery/CoalTipple · not measured

zAI-Skills

Development marketplace for AI-Toolkit, consulting-toolkit, real-estate-investment, career-coach, and plugin-forge plugins.

Aznatkoiny/zAI-Skills · not measured

subagent-model-routing

Claude Code subagent-model-routing plugin for non-Claude models — flat dispatch + DAG orchestration.

Buckeyes22/subagent-model-routing · not measured