agentleFS
Duckz289Duckz289/Pipeline-to-proposal-Plugin-STANDARD/skills/setup-product-context/SKILL.md

setup-product-context

Skill2 starsChanged 4 months agoMITView on GitHub

Saves what the user sells so every future email and proposal can reuse it. Triggered by "lưu context sản phẩm", "nhớ giúp mình bán gì", "cập nhật giá của mình", "save my offer", "remember what I sell", "update my pricing".

The file

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 3 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Ask one at a time, never as a form:

    line 15
  2. Distinctive:

    Only save what is real and the user would show a customer.

    line 53
  3. Distinctive:

    If the user enters something that reads like a result but has no source, ask one question:

    line 58

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “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

Save product context

Fill in once, use forever. Saved to <DATA>/product-context.md per ${CLAUDE_PLUGIN_ROOT}/policies/state-management.md.

Three questions is enough to start

Ask one at a time, never as a form:

  1. What do you sell? (one sentence is fine)
  2. Starting price, or a range?
  3. What kind of company is the best fit?

Save whatever they give. Leave the rest empty — the plugin will ask as needed.

Do not ask up front for: case studies, common objections, tone, default CTA, detailed packages, or the user's own name — sender_name is not part of the opening three questions. It is asked separately, later, per the "Sender name" rule in ${CLAUDE_PLUGIN_ROOT}/policies/safety-core.md (not before the second buyer-facing draft, and never invented as "Minh" or anything else in the meantime). When updating this file for any other reason, preserve an existing sender_name — do not drop it just because this particular save wasn't about it.

When to save automatically

product-context.md is empty and the user just stated their offer in any message → save it, no need to ask. Just confirm in one line.

File already has content → show a preview of what would change and ask before overwriting.

Output

### Đã lưu
Bán: phần mềm quản lý kho, từ 3 triệu/tháng
Khách hợp: công ty logistics, kho vừa và nhỏ

Còn thiếu: bằng chứng kết quả thật. Có thì gửi mình lưu thêm — email sẽ thuyết phục hơn.

Short. Do not read the whole file back to them. Headings follow the user's language.

The proof section

The most important and most error-prone part.

  • Only save what is real and the user would show a customer.
  • A desired outcome must be recorded as desired, not as achieved.
  • Empty beats invented — the plugin will not use what it does not have, but it also cannot verify what the user typed in.

If the user enters something that reads like a result but has no source, ask one question:

Con số này mình đã đo ở khách nào chưa, hay là mục tiêu mình hướng tới ạ?
Em hỏi để biết có nên đưa vào email gửi khách không.

Refresh prompts

File is 90+ days old → ask one question, not the whole list:

Thông tin sản phẩm mình lưu từ hồi tháng 4. Giá hay gói có gì đổi không ạ?

Also worth asking after a deal is lost for reasons about price or scope — a sign the saved context has drifted from reality.

References

  • references/product-context-fields.md — full structure when the user wants to fill it in properly
  • ${CLAUDE_PLUGIN_ROOT}/knowledge/product-context-template.md — empty template

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.