isolate-subagent-verification-from-live-worktree

isolate-subagent-verification-from-live-worktree is a skill for Claude Code from wan-huiyan/agent-traffic-control. It costs 342 tokens per session (1,567 once invoked), scanned A, original, MIT.

A verification procedure for running tests or checks in a clean Git worktree, an isolated checkout of a branch.

In plain words
What is it for?
Use it when assigning another agent to test a branch, especially during concurrent editing, staging, committing, or merging.
Why use it?
It prevents results from accidentally including the main agent’s uncommitted edits, which can make a branch appear verified when it was not.

Skill for Claude Code

Written for Claude Code: disable-model-invocation in frontmatter. Also seen: mentions CLAUDE.md; mentions subagents; mentions Claude Code.

Part of the agent-traffic-control plugin — 105 skills shipped together

Good fit Use it when assigning another agent to test a branch, especially during concurrent editing, staging, committing, or merging.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/wan-huiyan/agent-traffic-control/isolate-subagent-verification-from-live-worktree
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.

Any agent
npx skills add wan-huiyan/agent-traffic-control --skill isolate-subagent-verification-from-live-worktree
Clone the repo
git clone --depth 1 https://github.com/wan-huiyan/agent-traffic-control

Made for: Claude Code.

Or install agent-traffic-control, the plugin that ships this one along with the rest of its 105 skills.

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.

agentmods badge for isolate-subagent-verification-from-live-worktree

README.md
[![agentmods](https://agentmods.dev/badge/skills/wan-huiyan/agent-traffic-control/isolate-subagent-verification-from-live-worktree/github.svg)](https://agentmods.dev/skills/wan-huiyan/agent-traffic-control/isolate-subagent-verification-from-live-worktree)
Your own site
<a href="https://agentmods.dev/skills/wan-huiyan/agent-traffic-control/isolate-subagent-verification-from-live-worktree"><img src="https://agentmods.dev/badge/skills/wan-huiyan/agent-traffic-control/isolate-subagent-verification-from-live-worktree/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.

agentmods 80×15 button for isolate-subagent-verification-from-live-worktree

Your own site · 80×15
<a href="https://agentmods.dev/skills/wan-huiyan/agent-traffic-control/isolate-subagent-verification-from-live-worktree"><img src="https://agentmods.dev/badge/skills/wan-huiyan/agent-traffic-control/isolate-subagent-verification-from-live-worktree.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 342 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,567 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.

ModelPer sessionOnce invoked
Fable 5.1 $0.00342 $0.01567
Opus 5 $0.00171 $0.00783
Sonnet 5 $0.00068 $0.00313
Haiku 4.5 $0.00034 $0.00157

Measured 9d ago against content hash 4a2a7772b9e1, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-09, from the pricing page.

Security

Grade A, and why

isolate-subagent-verification-from-live-worktree 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 9d 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/agent-traffic-control/skills/isolate-subagent-verification-from-live-worktree/SKILL.md · 127 lines

How it starts

The opening of the file, as written. The whole thing — 127 lines — stays where its author put it; the contents beside it link to each section on GitHub.

A verifying agent in your live worktree measures your uncommitted work

Problem

You dispatch a subagent to verify a merge: "run the four suites and the validator against main, report the numbers." You point it at the worktree you are working in, because that is where the venv is symlinked and the tooling is set up.

Then you keep working — editing, staging, committing, merging PRs.

The agent's gate run measures the branch plus whatever is in your working tree at that instant. There is no clobber, no error, and nothing looks wrong: the suites really execute, the numbers it reports are real numbers, and it prints them with a commit SHA next to them. They are simply counts of a tree that exists only on your disk and will never be merged.

The report then becomes the session's evidence — pasted into a PR body, a handoff, a tracker card — as "verified green at <sha>".

Why this is worse than the clobber cases

The sibling failures in this collection are loud once you look:

  • dispatched-bash-agent-git-checkout-clobbers-uncommitted-edit — your edit vanishes.
  • concurrent-session-checkout-clobbers-shared-worktree — the branch changes under you.
  • subagent-read-stale-worktree-needs-head-pin — the agent's cited lines contradict yours.

Here nothing contradicts anything. The only signal is a total that does not reconcile with the same command run at a different moment — and totals move for legitimate reasons all day in a repo with parallel sessions, so the delta reads as normal churn.

Context / trigger conditions

  • You dispatched an agent to "verify", "run the gates", "confirm main is green", or "check the suite counts", and its prompt named a worktree path you are also using.
  • An agent's reported total disagrees with your own run of the same command, with no merge in between to explain it.
  • An agent mentions, in passing, that the working tree was modified with files it did not write, or that HEAD moved mid-run.
  • A git status in the agent's transcript is non-empty when you expected a clean tree.
  • You are about to publish "verified at <sha>" using numbers gathered while you had unstaged edits.

Read the full file on GitHub · 127 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. 9d ago First seen · 127 lines · 342 tokens per session scan A 4a2a7772b9e1

Subscribe to this mod's changes

isolate-subagent-verification-from-live-worktree is a skill published in the GitHub repository wan-huiyan/agent-traffic-control (3 stars, last pushed 4d ago), licensed MIT. It adds 342 tokens to every session and 1,567 once invoked, about $0.0017 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.