requirement-to-prd
Skill0 starsChanged 4 months agoMITView on GitHub
One-shot skill that reads raw requirement material and writes a finished PRD file to disk. Takes messy meeting notes, a client requirement document, an RFP, or a discovery-call transcript, analyzes it into structured requirement intelligence, then produces and saves a complete Product Requirements Document as a Markdown file. Use when the user hands over unstructured requirements and asks to "make a PRD", "write a PRD file", "generate a product requirements document", or "turn these notes into a PRD" and wants an actual file, not just text in the chat.
2 other projects use a version of this.
Adopted by
Most recently taken up by
- chenpianzhou/requirement-to-prd-skilltheir own version, first seen Jun 13, 2026
- Throokie/claude-code-skillstheir own version, first seen Mar 24, 2026
Other versions2 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 2
- Throokie/claude-code-skills★ 1same name
- chenpianzhou/requirement-to-prd-skill★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository6 other files it gives its agents.
The file
We split this skill into the rules it gives an agent
5 rules. About 0 are common, 2 are distinctive, and 3 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
3
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 21
Do not ask them to reformat it.
- Very specific: line 49
Use the exact structure in assets/prd-template.md.
- Very specific: line 61
Check the draft against references/prd-quality-checklist.md.
- Distinctive: line 73
After writing, confirm the saved path to the user and give a two-line summary (functional-requirement count, open-question count).
- Very specific: line 119
User wanted the file somewhere specific: If they named a path or filename, honor it.
≈ “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.
Files
Requirement to PRD
A self-contained shortcut: raw requirements in, a saved PRD.md out. Unlike the
two-stage requirement-analyzer → prd-generator path, this skill does the analysis
inline and its headline deliverable is a written file on disk, not chat text. Reach
for it when someone says "here are the notes, give me a PRD file."
Instructions
Step 1: Read the input as-is
Take whatever the user provides — notes, a document path, a transcript, or pasted text. Do not ask them to reformat it. If several sources are given, merge them. If they give a file path, read that file.
Step 2: Analyze the requirements (in your head, then on paper)
Extract the structured brief before writing any PRD prose. Pull out:
- Project name and client — literal names if stated; otherwise propose a short working name and add "confirm project name" to Open Questions.
- Objective — one outcome-focused sentence.
- Goals — measurable where the input supports it.
- Stakeholders — decision maker, budget holder, end users, integration owners.
- Functional requirements — one capability per line, verb-first.
- Non-functional requirements — the frequently-missed ones. Scan deliberately
against the checklist in
references/analysis-guide.md(performance, security, compliance, availability, scalability, usability/offline, localization, data restrictions). - Constraints — deadline, budget, mandated stack, integrations, data restrictions.
- Assumptions — anything you inferred; keep it visible.
- Open questions — specific, answerable gaps.
Apply the signal test: if removing a line does not change what gets built, drop it. "The client seemed excited" is not a requirement.
Step 3: Write the Project Vision
One paragraph naming who the product is for, the core value, and the outcome that defines success. This is the anchor for every scope decision below.
Step 4: Fill the PRD template
Use the exact structure in assets/prd-template.md. Fill every section from the brief
in Step 2. Where the input is genuinely silent, write "To be confirmed" and add the
item to Open Questions rather than inventing detail. Give every functional requirement a
stable ID (FR-01) and a MoSCoW priority; give every non-functional requirement an ID
(NFR-01).
Step 5: Make scope explicit
Populate Out of Scope deliberately — an empty Out of Scope is a red flag, not a convenience. State plainly what this release will not include; that is what prevents scope creep later.
Step 6: Run the quality gate
Check the draft against references/prd-quality-checklist.md. Every functional
requirement must trace to a goal or requirement in the input (no orphan features), and
nothing from the input may be silently dropped — anything cut appears in Out of Scope
with a reason. Fix failures before writing the file.
Step 7: Write the PRD file
Save the finished PRD as a Markdown file — this step is the point of the skill; do not stop at chat text.
- Filename: slugify the project name →
<project-slug>-prd.md(e.g.cod-reconciliation-prd.md). If the user named a path or filename, use theirs. - Location: the current working directory unless the user specified one.
- After writing, confirm the saved path to the user and give a two-line summary (functional-requirement count, open-question count).
Output structure
The written file ALWAYS follows this order (full template in assets/prd-template.md):
# [Project Name] — Product Requirements Document
## Project Vision
## Objectives and Success Metrics
## Personas
## Scope (In Scope)
## Out of Scope
## Functional Requirements
## Non-Functional Requirements
## Dependencies and Integrations
## Risks
## Open QuestionsExample
Input: messy last-mile-delivery discovery notes (rider app + merchant portal, COD reconciliation pain, offline-tolerant, cheap Android, "launch before Eid", no real customer data for testing).
Output: a saved cod-reconciliation-prd.md whose Vision reads —
"For a last-mile delivery operator whose finance team loses half a week reconciling cash-on-delivery collections by hand, this platform automates COD reconciliation and merchant settlement: riders record collected cash on a resilient mobile app and merchants track parcel status and settlement timing in a portal, so merchants are paid accurately in days sooner and disputes fall."
Out of Scope explicitly excludes in-app customer payments and a public consumer app for this release; Open Questions capture the undecided Eid deadline and the bKash-vs-bank settlement choice.
Troubleshooting
- PRD is vague. The input brief was thin. Do not pad — surface the gaps as Open Questions and tell the user what to confirm.
- PRD invents features nobody asked for. Cut anything not traceable to a goal or requirement in the input. Every PRD line traces back to the requirement that justifies it.
- User wanted the file somewhere specific. If they named a path or filename, honor
it; otherwise default to
<project-slug>-prd.mdin the working directory and tell them where it landed.
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
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.

