NewsBlur is a personal RSS news reader and social network that displays stories, the original sites they come from, and tools for filtering and discussing them. Readers use it to follow subscriptions, train story preferences, search and tag articles, share reading, and optionally self-host the service. The catalogue entries provide commands, skills, instructions, hooks, and settings for working with the NewsBlur project.
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.
npx agentmods add commands/samuelclay/newsblur/featuregit clone --depth 1 https://github.com/samuelclay/NewsBlurWrote 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/commands/samuelclay/newsblur/feature)<a href="https://agentmods.dev/commands/samuelclay/newsblur/feature"><img src="https://agentmods.dev/badge/commands/samuelclay/newsblur/feature.svg" alt="Measured on agentmods" 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 | $0.00012 | $0.00981 |
| Opus 5 | $0.00006 | $0.00491 |
| Sonnet 5 | $0.00002 | $0.00196 |
| Haiku 4.5 | $0.00001 | $0.00098 |
Grade A, and why
feature 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 5d 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 — 89 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Context
- Recent blog posts: !
ls -t blog/_posts/*.md | head -10 - Current features on prod: !
./utils/ssh_hz.sh -n happ-web-01 "docker exec -t newsblur_web python manage.py shell -c \"from apps.reader.models import Feature; [print(f'{f.date.strftime(\\\"%b %d, %Y\\\")}: {f.description}') for f in Feature.objects.all()[:6]]\""
Writing style
All text you write (feature descriptions and tweets) must sound like a human wrote it. Avoid AI slop:
- No em dashes or double hyphens -- don't use dashes as punctuation at all. Restructure the sentence instead.
- No overly polished or corporate language -- keep it direct and natural
- No filler phrases like "excited to announce", "we're thrilled", "game-changer"
- Write the way a developer would describe their own product
Your task
Create a new Feature entry on the NewsBlur production database for the topic: {{ arguments }}, then write tweets.
Step 1: Find the matching blog post
Search blog/_posts/ for a blog post matching the topic "{{ arguments }}". Read the full blog post to understand:
- What the feature does
- Key capabilities and highlights
- Which subscription tiers have access
- The blog post date and URL slug
Derive the blog URL from the filename: https://blog.newsblur.com/YYYY/MM/DD/slug-from-filename/
(e.g., blog/_posts/2026-03-25-daily-briefing.md → https://blog.newsblur.com/2026/03/25/daily-briefing/)
Step 2: Study existing feature style
Look at the existing features from the Context section above. Note:
- Length: 1-3 sentences, concise and punchy
- Tone: Product-focused, highlights what's new and useful
- Structure: Feature description +
<a href="...">Read the blog post</a>. - Content: Leads with what it does, packs in key details, sometimes mentions tier availability
- The description field contains raw HTML (the blog link is an anchor tag)
Step 3: Write 4-5 feature options
Write 4-5 different feature descriptions, each following the existing style. Vary them by:
- Which aspects of the feature to highlight (breadth vs. depth)
- Length (shorter vs. slightly longer)
- What details to include (sections, delivery methods, customization, tier info)
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.
- 5d ago First seen · 89 lines · 12 tokens per session scan A 6fe90881a7ec
feature is a command published in the GitHub repository samuelclay/NewsBlur (7,608 stars, last pushed yesterday), licensed MIT. It adds 12 tokens to every session and 981 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.
Other commands, from other repositories
release
릴리즈 절차 — 버전 추천부터 릴리즈 PR까지 (changesets, 가이드형).
compound
현재 작업에서 재사용 패턴을 추출해 AGENTS.md·문서를 업데이트한다. 같은 종류의 버그·구현을 2회 이상 해결했거나 작업 종료 시 재사용 가치가 보일 때 사용. 일회성 작업에는 쓰지 않는다.
qa
QA 전문가 — 테스트 기획 및 작성. 새 기능에 테스트가 필요하거나 기존 테스트의 신뢰성(Potemkin·Flaky)이 의심될 때 사용. 실제 기능·성능 검증 중심.
audit
Scan a codebase for existing observability instrumentation and identify gaps.
mobile-verify
Run automated verification loops with pass@k metrics for mobile testing. Executes tests multiple times to detect flakiness.
feature-learn
View and manage patterns learned from feature builds. Shows extracted instincts, confidence scores, and feature completeness history.