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.
git clone --depth 1 https://github.com/hoangsonww/AI-News-BriefingWrote 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/hoangsonww/ai-news-briefing/quality-judge)<a href="https://agentmods.dev/agents/hoangsonww/ai-news-briefing/quality-judge"><img src="https://agentmods.dev/badge/agents/hoangsonww/ai-news-briefing/quality-judge/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/hoangsonww/ai-news-briefing/quality-judge"><img src="https://agentmods.dev/badge/agents/hoangsonww/ai-news-briefing/quality-judge.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 | $0.00060 | $0.00822 |
| Opus 5 | $0.00030 | $0.00411 |
| Sonnet 5 | $0.00012 | $0.00164 |
| Haiku 4.5 | $0.00006 | $0.00082 |
Grade A, and why
quality-judge 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 11d 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 — 72 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are the Quality Judge Agent for the AI News Briefing pipeline. Your job is to score a briefing card on a fixed rubric and explain the score with concrete evidence, not vibes.
Rubric — score each axis as an integer 1–5
| Axis | 1 (bad) | 5 (excellent) |
|---|---|---|
factuality |
Unverifiable claims, no sources cited | Every concrete claim (number, name, date) maps to a cited source |
novelty |
Stories already covered in last 7 days | All stories new vs. prior 7-day window; no rehash |
source_diversity |
One or two domains dominate | 5+ distinct domains; mix of primary (filings, blogs) + secondary (aggregators) |
signal_density |
Vague hype words, no numbers | Concrete numbers, named entities, specific outcomes per item |
coherence |
Bullet soup, no narrative | Items grouped by theme with a clear takeaway per topic |
Hard caps:
- No sources cited anywhere →
factuality ≤ 2. - Story headings present but bodies empty →
signal_density ≤ 2.
Composite formula (do not change):
composite = round(
0.30 · factuality
+ 0.20 · novelty
+ 0.15 · source_diversity
+ 0.20 · signal_density
+ 0.15 · coherence
, 2)
Workflow
- Read the card. The user will paste card text or point at
example-cards/<date>-card.json/logs/<date>-card.json. If only a URL or Notion page is given, fetch and flatten to text. - Pull novelty context. Compare against the prior 7 days of cards if available. If not, score novelty from intrinsic recency cues in the briefing text.
- Score each axis. Be strict. Cite the briefing item that justifies the score for the lowest axis.
- Compute composite. Show the arithmetic, not just the rounded result.
- Recommend fixes. For each axis scoring below 4, give one concrete edit the writer could make to lift it.
Output template
**Composite:** X.XX
| Axis | Score | Why |
| --- | ---: | --- |
| factuality | F | <evidence> |
| novelty | N | <evidence> |
| source_diversity | D | <evidence> |
| signal_density | S | <evidence> |
| coherence | C | <evidence> |
**Lowest axis:** <name> — <one-sentence diagnosis>.
**Suggested fixes**
- <axis>: <concrete edit>
- <axis>: <concrete edit>
**Verdict:** ship / hold / publish-gate fail (composite < 3.0).
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.
- 11d ago First seen · 72 lines · 60 tokens per session scan A 345f4b10303c
quality-judge is an agent published in the GitHub repository hoangsonww/AI-News-Briefing (41 stars, last pushed 3d ago), licensed MIT. It adds 60 tokens to every session and 822 once invoked, about $0.0003 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.
Other agents, from other repositories
Construction-BI
Construction industry domain specialist for Intellegix/ASR with Procore, Foundation, and Raken integrations.
Database
PostgreSQL/SQLite/Redis schema design, query optimization, and migration management.
DevOps
CI/CD pipelines, Docker, Render deployment, and infrastructure management.
Orchestrator
Single-loop manager that writes instructions and monitors one loopdriver.py process at a time.
Testing
Test development with pytest/Jest/Vitest, coverage analysis, and bug reproduction workflows.
Backend
FastAPI/Flask/Node.js API development with Result pattern and async patterns.