An AI review step for risky operations such as production database changes, releases, mass deletion, or pushing to a protected branch. It returns a go-or-hold decision with reasons after checking the proposed action.

Use it before deploying, publishing, running database migrations, deleting many records or files, changing all-environment settings, or pushing to main or a release branch.

It adds a deliberate safety check before an operation that could affect production, many users, or valuable data. The review considers intent, changed files, impact, reversibility, and prerequisites.

llm-gate — Command for Claude Code by atuljha23

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/atuljha23/holocron

Made for: Claude Code.

Or install holocron, the plugin that ships this one along with the rest of its 11 skills, 24 commands, 14 agents, 6 hooks.

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/commands/atuljha23/holocron/llm-gate.svg)](https://agentmods.dev/commands/atuljha23/holocron/llm-gate)

Your own site

<a href="https://agentmods.dev/commands/atuljha23/holocron/llm-gate"><img src="https://agentmods.dev/badge/commands/atuljha23/holocron/llm-gate.svg" alt="Measured on agentmods" height="20"></a>

Per session 48 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.

When invoked 577 The whole file, excluding the scripts and references it only reads on demand.

Security scan C 1 finding. 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	$0.00048	$0.00577
Opus 5	$0.00024	$0.00289
Sonnet 5	$0.00010	$0.00115
Haiku 4.5	$0.00005	$0.00058

Measured 7d ago against content hash c5dfb5d8f582, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-07, from the pricing page.

Security

Grade C, and why

llm-gate scanned grade C 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 7d 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.

Recursive force deletehighDestructive command

rm -rf with a variable or a broad path is one typo away from removing the wrong tree.

- Before a mass delete (`rm -rf`, `DELETE FROM`, `kubectl delete`).

commands/llm-gate.md · 66 lines

What it actually says

/holocron:llm-gate

Ask: is this about to do what I think it's about to do?

When to invoke

Before git push to main / master / a release branch.
Before running a production migration.
Before a mass delete (rm -rf, DELETE FROM, kubectl delete).
Before publishing a release (npm publish, gh release create).
Before a config change that crosses environments (rotating a secret, flipping a feature flag for all tenants).

Proposed operation

$ARGUMENTS

Checklist the gate walks

Before returning a verdict, satisfy these:

Intent clear. State in one sentence what you believe will happen.
Diff surveyed. If it's a code change, git diff --stat and git diff on the relevant range. If it's a DDL, inspect the migration file.
Blast radius. Who/what is affected? One tenant or all? One service or the whole cluster? One row or a table?
Reversibility. Can this be undone in < 5 minutes? < 1 hour? Not at all?
Preconditions. Has the quality gate run (tests, lint, typecheck, staging verification)? Is on-call available?
Suspicious patterns. Force-push to a protected branch, secrets in the diff, migration without a DOWN, delete without a WHERE.

Verdict

## Gate verdict: GO  |  HOLD  |  BLOCK

### Intent
<one-sentence read of what's about to happen>

### Blast radius
<scope>

### Reversibility
<minutes / hours / not>

### Concerns
- <item>  (severity: low/medium/high)
- <item>
- ...

### Required before GO (if HOLD)
- <action>
- <action>

### Why BLOCK (if BLOCK)
<one or two sentences — specific, actionable>

Do not

Do not approve your own work without reading the diff.
Do not rubber-stamp with "LGTM" — produce concerns, even minor ones, so the user sees them.
Do not block on cosmetic issues. BLOCK is for "this will break prod".

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.

7d ago First seen · 66 lines · 48 tokens per session scan C c5dfb5d8f582

Subscribe to this mod's changes

llm-gate is a command published in the GitHub repository atuljha23/holocron (2 stars, last pushed 4mo ago), licensed MIT. It adds 48 tokens to every session and 577 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it C with 1 finding (recursive force delete). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Claude Code Security agent-orchestration ai-agents ai-coding ai-workflow claude

llm-gate

Getting it into your agent

Wrote this? Show the measurements

What it costs to keep this loaded

Grade C, and why

Recursive force deletehighDestructive command

What it actually says

/holocron:llm-gate

When to invoke

Proposed operation

Checklist the gate walks

Verdict

Do not

What this file has done since we first saw it

Other commands, from other repositories

learn

insights

wiki

handoff

commit

context-optimizer