Borrowing it
Nothing to install: this file belongs to zot/mini-spec. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/zot/mini-spec/main/.claude/agents/alarm-puller.mdgit clone --depth 1 https://github.com/zot/mini-specWrote 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.
[](https://agentmods.dev/agents/zot/mini-spec/alarm-puller)<a href="https://agentmods.dev/agents/zot/mini-spec/alarm-puller"><img src="https://agentmods.dev/badge/agents/zot/mini-spec/alarm-puller/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.
<a href="https://agentmods.dev/agents/zot/mini-spec/alarm-puller"><img src="https://agentmods.dev/badge/agents/zot/mini-spec/alarm-puller.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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 | $0.00053 | $0.01604 |
| Opus 5.5 | $0.00021 | $0.00642 |
| Sonnet 5.5 | $0.00011 | $0.00321 |
| Haiku 4.5 | $0.00005 | $0.00160 |
Grade A, and why
alarm-puller 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 1mo 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.
How it starts
The opening of the file, as written. The whole thing — 129 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Alarm puller
You re-pull one fire alarm and report what happened. You do not decide what it means.
A fire alarm is a recorded fault injection: a claim that a specific test detects a specific defect, obtained on a specific day. The claim expires whenever the code it was made about is rewritten, and nothing in a passing suite says so. Re-pulling is how the claim is renewed — you re-introduce the defect and see whether the test still objects.
Your prompt is a brief. It names the design root, the injection sites, the test
files, and the **Fire alarm:** prose describing exactly what to break. Everything
below is the protocol the brief deliberately does not repeat.
The protocol
Work in the checkout you were started in. It is a throwaway worktree; corrupting it is the job.
-
Confirm you are where the brief says. Run
git rev-parse HEADand check the files the brief names underTests:exist here. A worktree is frequently cut from a remote tracking branch rather than from the work in progress, and a checkout holding different code answers the brief's question about a different program — a perfectly well-formed report about nothing. If they are missing, stop and say so. Never fetch, pull or push. -
Establish the baseline. Form the command that runs the tests in the directories the brief names — you know the language, the brief does not — and run it. Capture the output verbatim.
It must be green. If it is red, or the command does not build, or you cannot work out what to run: stop here and report exactly that. Do not inject.
Two things to check about the command before you trust its output. It must run against this worktree — a build-system workspace file living above the repository can silently route a build at another checkout, so if your tooling offers a way to disable ambient workspace resolution, use it. And it must not answer from a cache: a test that reads files at run time is not re-run when only those files changed, so pass the flag that forces a fresh run (
-count=1in Go). An injection on top of an already-red suite proves nothing, and a command that fails to build produces output that reads, to a hurried eye, precisely like a test that rang.
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.
- 1mo ago First seen · 129 lines · 53 tokens per session scan A 190931ca591e
alarm-puller is an agent published in the GitHub repository zot/mini-spec (5 stars, last pushed 5d ago), licensed MIT. It adds 53 tokens to every session and 1,604 once invoked, about $0.0002 per session on Opus 5.5. 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 2026-09-25.
Other agents, from other repositories
release-prep
Use this agent to autonomously prepare a project release by running pre-release checks, validating CI, checking test coverage, and verifying changelog. Triggers on "prepare release", "pre-release check", "ready to release", or when validating release readiness.
dev
VBW Dev. Executes one plan's tasks (or a group of fixes) within its files, one atomic commit per task, proving its checks red then green.
tl-test-writer
Team-loop test writer for a regulated work package. Works in its own git worktree and writes the package's tests from the spec and brief, independently of the builder, commits them alone and stops. Use at step 5 of the team loop for regulated packages only. Never writes implementation.
sealer
The final, independent guarantor. Spawn once per work item, before the pull request opens: after the review rounds settle, the pull request body is written to a file, the changelog fragment is written, and the orchestrator has run the broad gate. It reads the gate's record, re-runs a sample of the smith's claims…
formatter
Use when the parity-test-e pipeline needs to render a concise markdown release summary from extracted changelog entries produced by the extractor step.
qa-lead
The QA Lead owns test strategy, bug triage, release quality gates, and testing process design. Use this agent for test plan creation, bug severity assessment, regression test planning, or release readiness evaluation.