agentleFS
XuHo-ITXuHo-IT/Content-Agent-Kit/skills/freshness-check/SKILL.md

freshness-check

Skill8 starsChanged 4 months agoMITView on GitHub

Check and renew the facts the kit relies on that go stale — platform upload limits, Vietnamese advertising and content law, platform policy news (monetisation, AI-content labels), skill licences, the template registry, Vietnamese speech models — by researching the latest official sources on the web. Use at the start of any session that will publish, advertise or deliver to a client, when another skill says to run it first, or when the user asks "có chính sách gì mới", "kiểm tra cập nhật", "luật mới về quảng cáo", "TikTok đổi gì chưa".

3 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

5 rules. 0 are common, 4 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 3 edited versions we read keep these.

  • 4

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Research and code items: search with every query in queries (they already carry this year and month).

    line 39
  2. Distinctive:

    Check the date ON the source page: an undated or old page does not prove a current rule.

    line 42
  3. Distinctive:

    If it changes what may be posted (e.g.

    line 55
  4. Distinctive:

    Never weaken a rule on a news article alone.

    line 64
  5. Distinctive:

    Every changed number keeps its source and the date it was read.

    line 66

“Common” is measured: kept by at least 6 in 10 of the 3 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Freshness check — kiểm và cập nhật những gì kit dựa vào

Hướng dẫn từng bước cho người dùng

Chạy lệnh node scripts/… ở thư mục gốc của kit — thư mục có AGENTS.md, hai cấp trên file này (nếu skill được cài bằng install-skills.mjs --kit, đường dẫn nằm trong file .kit-root cạnh file này).

Kit dựa vào những thứ nó không kiểm soát: TikTok nhận video dài bao nhiêu, luật quảng cáo cho giảm giá bao nhiêu phần trăm, nền tảng có bắt gắn nhãn "do AI tạo" không. Mỗi điều chỉ đúng ở thời điểm được viết. Skill này tìm ra điều nào đã cũ, lên mạng tra bản mới nhất từ nguồn chính thức, rồi cập nhật lại, có ghi nguồn.

Dẫn người dùng theo đúng các bước dưới, bằng lời đơn giản. Đầu mỗi bước báo "Bước n/N — <tên bước>". Thường mất 2–10 phút, tuỳ có bao nhiêu mục đã cũ.

BướcAgent làm gìNgười dùng cần làm
1. Xem mục nào đã cũnode scripts/maintenance/freshness.mjs --json --stale—
2. Chạy phần tự độngMục auto: chạy lệnh được ghi sẵn (licence-watch.mjs, registry-watch.mjs)—
3. Tra cứuMục research / code: tìm bằng các câu truy vấn được ghi sẵn, đọc nguồn chính thức—
4. Cập nhậtCó thay đổi → sửa đúng file được chỉ, ghi nguồn; không đổi → --mark <id> --note "…"Duyệt thay đổi về luật trước khi áp dụng
5. Báo lạiTóm tắt điều gì mới, điều gì ảnh hưởng tới video/bài sắp đăngĐọc tóm tắt

Khi nào chạy: đầu mỗi phiên sẽ đăng bài, chạy quảng cáo, hoặc giao video cho khách. daily-run, video-and-post và ads-policy-check đều gọi skill này ở bước đầu. Nếu không có mục nào cũ thì bỏ qua trong vài giây, không làm chậm gì.

Quy trình chi tiết (cho agent)

  1. List. node scripts/maintenance/freshness.mjs --json --stale. Nothing stale → say so in one line and return to the calling skill.
  2. Auto items (kind: "auto"): run command. Exit 1 = drift: fix as update says (licences: edit skills/registry.json; after READING a licence GitHub cannot classify, licence-watch.mjs --pin <id>). Then --mark <id>.
  3. Research and code items: search with every query in queries (they already carry this year and month). Prefer the domains in official; a news article is a lead, not a source — follow it to the platform's help page or the legal text (vbpl.vn for Vietnamese law). Check the date ON the source page: an undated or old page does not prove a current rule.
  4. Decide, per item:
    • Changed and it affects a kit rule (kind: "code"): edit the file in file as update says, bump its CHECKED date, cite the source URL in a comment or sources. Law (vn-ads-law, vn-content-law): show the user the text and the proposed change and wait for a yes — a wrong legal rule blocks or permits real ads.
    • Changed, no rule in code (kind: "research"): append to brain/policy-news.md:
      ## 2026-11-11 — <item id>
      - **Gì đổi:** … **Từ khi nào:** … **Ai bị ảnh hưởng:** …
      - **Kit cần làm gì:** … (or "không")
      - Nguồn: <url> (ngày trên trang: …)
      If it changes what may be posted (e.g. a mandatory AI label), also say it in the review-gate notes for this session.
    • Nothing changed: node scripts/maintenance/freshness.mjs --mark <id> --note "<what you checked>".
  5. Report in 3–6 lines: what is new, what it means for today's content, what you changed. Unsure (sources disagree, no official page)? Say so — do not resolve it by guessing; record it as a warning, the way delivery-specs.mjs keeps disputed limits soft.

Rules

  • Never weaken a rule on a news article alone. Tightening on a credible report is fine (warn); loosening needs the official text.
  • Every changed number keeps its source and the date it was read.
  • state/freshness.json and brain/policy-news.md are this project's own record (gitignored). A fact that changes the KIT for everyone (a platform limit, a licence) belongs in a commit to the kit, so the next user starts fresh.
  • Items, their max ages and queries: scripts/maintenance/lib/freshness.mjs. Add one when the kit starts relying on a new outside fact.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.