kube-agents: Command for Claude Code

.claude/commands/pr-review-batch.md

pr-review-batch is a command for Claude Code from gke-labs/kube-agents. It costs 21 tokens per session (9,186 once invoked), scanned A, original, Apache-2.0.

A command for reviewing one or more GitHub pull requests, which are proposed code changes, in separate temporary workspaces. It performs an initial check and then assigns each eligible pull request to its own review process.

In plain words
What is it for?
Use it to review pull requests in parallel, check their pre-review status, inspect changes from the correct branch or commit, and receive a concise structured result for each one.
Why use it?
It keeps reviews isolated from one another and verifies review claims before reporting results. The initial check also determines whether a pull request has already been covered and what review starting point to use.

Command for Claude Code

Written for Claude Code: argument-hint in frontmatter. Also seen: mentions subagents; mentions AGENTS.md.

This is gke-labs/kube-agents's own configuration. It tells Claude Code how to work on kube-agents itself, so it is not a mod to install elsewhere. Copy it as a starting point and replace the rules that are about this project. Everything kube-agents configures →

Reuse

Borrowing it

Nothing to install: this file belongs to gke-labs/kube-agents. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file
curl -O https://raw.githubusercontent.com/gke-labs/kube-agents/main/.claude/commands/pr-review-batch.md
Clone the repo
git clone --depth 1 https://github.com/gke-labs/kube-agents

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.

agentmods badge for pr-review-batch

README.md
[![agentmods](https://agentmods.dev/badge/commands/gke-labs/kube-agents/pr-review-batch/github.svg)](https://agentmods.dev/commands/gke-labs/kube-agents/pr-review-batch)
Your own site
<a href="https://agentmods.dev/commands/gke-labs/kube-agents/pr-review-batch"><img src="https://agentmods.dev/badge/commands/gke-labs/kube-agents/pr-review-batch/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 pr-review-batch

Your own site · 80×15
<a href="https://agentmods.dev/commands/gke-labs/kube-agents/pr-review-batch"><img src="https://agentmods.dev/badge/commands/gke-labs/kube-agents/pr-review-batch.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 21 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 9,186 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 $0.00021 $0.09186
Opus 5 $0.00010 $0.04593
Sonnet 5 $0.00004 $0.01837
Haiku 4.5 $0.00002 $0.00919

Measured 6d ago against content hash 148340a7b662, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-09, from the pricing page.

Security

Grade A, and why

pr-review-batch 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 6d 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/pr-review-batch.md · 618 lines

How it starts

The opening of the file, as written. The whole thing — 618 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Review these pull requests in gke-labs/kube-agents: $ARGUMENTS

PRs always live in that repo. Everything else — remote names, the base branch, the checkout location — is discovered at runtime, so this command works for any teammate in any clone regardless of what they called their remotes or where they cloned to.

Run Phase −1 in the main loop first, for every PR number, and wait for my answer. Then spawn one subagent per PR number that survives it, all in a single message so they run concurrently. Each subagent owns exactly one PR end to end and reports back a short structured result. Do not review any PR yourself in the main loop — your job is to pre-flight, fan out, then relay.

Give each subagent everything under Subagent instructions (per PR) verbatim, with <N> replaced by its PR number, plus the PR's pre-flight verdict, the review mode I chose for it, and the SHA that mode starts from when it is a narrowed one. Phase −1 is yours and stops at that heading: it asks me a question, which a subagent cannot do, so handing it on would either hang the subagent or have it answer on my behalf.


Phase −1 — Pre-flight: is this review already covered? (main loop)

Every PR here is read by kube-agents-bot on the way in, and the repo requires the author to have run review-adversarial over their own diff and written the disposition into the body before opening. When both of those happened and neither has gone stale, a third hostile read usually buys nothing — so find that out before spending it.

This phase runs in the main loop, before any subagent exists, for two reasons. Its output is a question for me, and a subagent has no way to ask one. And it is pure GitHub API — no worktree, no fetch, no diff read — so it costs two calls per PR, plus one per merge sitting after the bot's review, which is nearly always none or one.

The two queries

The repo is named literally in both: Phase −1 runs before Phase 0 defines $REPO.

Read the full file on GitHub · 618 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. 6d ago Changed · +9 lines 148340a7b662
  2. 10d ago First seen · 609 lines · 21 tokens per session scan A b1daa2685b55

Subscribe to this mod's changes

pr-review-batch is a command published in the GitHub repository gke-labs/kube-agents (53 stars, last pushed yesterday), licensed Apache-2.0. It adds 21 tokens to every session and 9,186 once invoked, about $0.0001 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 2026-08-30.