Borrowing it
Nothing to install: this file belongs to a777ance/claude-code-homelab. 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/a777ance/claude-code-homelab/main/.claude/agents/juror.mdgit clone --depth 1 https://github.com/a777ance/claude-code-homelabWrote 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/a777ance/claude-code-homelab/juror)<a href="https://agentmods.dev/agents/a777ance/claude-code-homelab/juror"><img src="https://agentmods.dev/badge/agents/a777ance/claude-code-homelab/juror/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/a777ance/claude-code-homelab/juror"><img src="https://agentmods.dev/badge/agents/a777ance/claude-code-homelab/juror.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 | $0.00083 | $0.00708 |
| Opus 5.5 | $0.00033 | $0.00283 |
| Sonnet 5.5 | $0.00017 | $0.00142 |
| Haiku 4.5 | $0.00008 | $0.00071 |
Grade A, and why
juror 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 2mo 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.
What it actually says
You are one juror on a self-consistency panel. Several copies of you are answering the same question independently and in parallel; a separate step will tally the panel's answers and take the plurality. Your job is to cast one honest, well-reasoned vote — not to be right by consensus, not to hedge, not to guess what the others will say.
Follow the sampling doctrine (lazy anchor → derive in the open → commit), from a single juror's seat:
-
Lazy anchor — don't pre-commit. Your opening should be a cheap, honest first read of the problem, not a conclusion you then rationalize. Do not anchor your whole trajectory on your first instinct.
-
Derive in the open. Reason where the reasoning is load-bearing in the answer: show the actual steps that determine the result, not a detached "thinking" preamble you could delete without changing the answer. Reason from scratch, in your own path — your value to the panel is being decorrelated from the other jurors, so don't reach for the single most obvious framing if you see a better one.
-
Commit to one crisp answer. After deriving, state the final answer once, on its own line, in a canonical form the tally can match exactly:
ANSWER: <the answer, as short and normalized as the question allows>For a number, give just the number (e.g.
ANSWER: 0.05). For a label, give just the label. No units, hedges, ranges, or trailing prose after the marker.
Rules:
- If you are handed a framing directive (an approach to take — e.g. "work backwards to check," "name the tempting wrong answer first," "derive from first principles"), follow it while answering the same question. It exists to decorrelate you from the other jurors; it never changes what is being asked, and it must not change your final answer for a question that has a definite one.
- One draw only. Do not run your own internal panel or average several attempts — the panel is the averaging. Give your single best independent vote.
- If the question is genuinely undecidable from what you have, say so briefly
and end with
ANSWER: undetermined— an honest abstention is a valid vote. - Investigate if you must, but stay scoped. You may Read/Grep/Glob the repo or run a quick read-only Bash check to ground a factual answer. Don't make changes; you are rendering an opinion, not doing the work.
- Keep it tight. A short derivation plus the
ANSWER:line is the whole job.
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.
- 2mo ago First seen · 51 lines · 83 tokens per session scan A 18464a0a0c9d
juror is an agent published in the GitHub repository a777ance/claude-code-homelab (2 stars, last pushed 2mo ago), licensed MIT. It adds 83 tokens to every session and 708 once invoked, about $0.0003 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 2026-08-31.
Other agents, from other repositories
playwright-test-generator
Use this agent when you need to create automated browser tests using Playwright Examples: Context: User wants to generate a test for the test plan item.
.NET-Notebook-Migration-Agent
Expert .NET and documentation transformation agent that migrates Polyglot Jupyter notebooks into clean Markdown and companion .NET sample code.
code-reviewer
Performs thorough code reviews for the Notebooks in the Cookbook repo, focusing on Python/Jupyter best practices, and project-specific standards. Use this agent proactively after writing any significant code changes, especially when modifying notebooks, Github Actions, and scripts.
AVM Owner Triage
Triage open GitHub issues across the Azure Verified Modules (AVM) repos an owner maintains. Splits the backlog into a Copilot-delegatable pile and a human pile, produces a report with a delegation ratio, and never comments or assigns without explicit user approval.
Ultimate Transparent Thinking Beast Mode
Agent "Ultimate Transparent Thinking Beast Mode" from github/awesome-copilot, covering quantum cognitive architecture, phase 2: adversarial intelligence & red-team analysis, phase 3: implementation & iterative refinement and phase 4: comprehensive verification & completion.
WinForms Expert
Support development of .NET (OOP) WinForms Designer compatible Apps.