Skip to content

PRD

Used in the playbook Project foundation, stage 3. Also used in Requirements specification (PRD or SRS)

Download prd.md — the source, as in the repository.

How to use

The template's own instructions:

Template from the UASK Engineering Hub, version 2, playbook mgmt.project-foundation, stage 3. Save as project/PRD.md. The front matter says which tier the project is on. A section opening with "Required: <tiers>" is required at those tiers only; "Optional" means never required; no marker means required at every tier. Delete these comments when a section is written. The earlier version is templates/prd-v1.md; a PRD without `template:` is checked against it and gets a warning.

Fields

Front-matter fields and sections of the document the template creates. A section marked Optional in the template may be deleted from a copy; every other field and section stays (tools/check_artifacts.py checks this for the kinds it knows).

Field or sectionWhereRequired
templatefront matterrequired
titlefront matterrequired
statusfront matterrequired
versionfront matterrequired
datefront matterrequired
ownerfront matterrequired
tierfront matterrequired
Problemsectionrequired
Users and contextsectionoptional
Scenariossectionrequired
Functional requirementssectionrequired
Non-functional requirementssectionrequired
Acceptance criteriasectionrequired
Constraints and assumptionssectionoptional
Parity with the current systemsectionoptional
External interfaces and environmentsectionrequired
Operabilitysectionrequired
Sources and traceabilitysectionoptional
Out of scopesectionrequired
Open questionssectionoptional

Example

A filled-in copy for an invented product, templates/examples/prd.md, in the Hub repository.It shows the level of detail to aim for, not text to reuse.

example-prd.md
---
template: prd/2
title: Product Requirements
status: draft
version: 0.1.0
date: 2026-03-01
owner: Product Owner
tier: standard
---
# PRD — Fieldnotes
Keywords MUST, SHOULD, MAY follow [RFC 2119](https://www.rfc-editor.org/rfc/rfc2119) /
[RFC 8174](https://www.rfc-editor.org/rfc/rfc8174). Ids (`S-…`, `FR-n`, `NFR-n`, `AC-n`; outcomes are `O-n` in
the vision) are stable, never reused, and referenced from the blueprint, roadmap, issues and tests.
## Problem
Field researchers lose notes when a cloud service cannot sync. Today they copy text between apps by hand, which costs about
an hour a week and sometimes loses a day of work.
## Users and context
Researchers on phones and laptops, often without a network for days.
## Scenarios
| ID | Scenario | Actor | Inputs | Result | Outcome |
|---|---|---|---|---|---|
| S-A | Write notes for a day without a network | Researcher | text, photos | notes saved on the device | O-1 |
| S-B | Merge a day of notes with the team's copy | Researcher | a Git host account | one history, no lost note | O-1 |
| S-C | Leave with every note | Researcher | none | a zip of Markdown files | O-2 |
## Functional requirements
| ID | Requirement | Scenario | Priority | State |
|---|---|---|---|---|
| FR-1 | The app MUST save a note to a file on the device without a network. | S-A | must | not started |
| FR-2 | The app MUST push and pull notes to a Git host when a network is present. | S-B | must | not started |
| FR-3 | The app MUST export all notes as Markdown files in one archive. | S-C | must | not started |
## Non-functional requirements
| ID | Characteristic | Requirement | Checked by |
|---|---|---|---|
| NFR-1 | Reliability | No note is lost when the app is killed while saving. | kill test in `tests/` |
| NFR-2 | Performance | A note of 10 000 characters opens in under 300 ms on the reference phone. | benchmark in CI |
## Acceptance criteria
- [ ] AC-1 FR-1 and NFR-1: a note written in flight mode is on disk after the app is killed.
- [ ] AC-2 FR-2: two devices edit different notes offline; after sync both have both.
- [ ] AC-3 FR-3: the archive opens in any text editor and holds every note.
## Constraints and assumptions
Assumes the user has a Git host account (assumption A-1, to be confirmed with testers).
## Parity with the current system
| Capability | State | Evidence |
|---|---|---|
| Write a note | not started | — |
## External interfaces and environment
| Interface or setting | Today | After the switch | Who owns it | Covered by |
|---|---|---|---|---|
| Git host | none | HTTPS remote chosen by the user | the user | FR-2 |
## Operability
| Point | Requirement (value) | Checked by |
|---|---|---|
| Availability | works fully offline; sync best effort | AC-1 |
| Recovery point (RPO) | no loss of a saved note | NFR-1 |
| Recovery time (RTO) | not applicable, no server | — |
| Restore test | before every release, by the lead; last on 2026-05-20 | release checklist |
| Rollback | the previous app version is reinstalled; notes stay as files | release checklist |
| Monitoring | crash reports opt-in; the lead reads them weekly | — |
| Who runs it | the owner; the lead covers | — |
| End of life | the runtime is supported until 2028-10; review in 2028-04 | — |
## Sources and traceability
| Requirement | Source |
|---|---|
| the note file (FR-1) | tester interviews, `research/2026-02-testers.md` |
## Out of scope
- Real-time collaborative editing.
## Open questions
- Which Git hosts must be supported in 1.0? Answers: the owner, by 2026-03-15; default: any HTTPS remote.

Completion checks

A copy is complete when:

  • every required field and section above is present, in the same order, and no bracketed placeholder is left;
  • the template's comments are deleted from the copy;
  • python3 .uask/hub/tools/check_artifacts.py . reports no failure for it: front matter, sections, ids, leftover placeholders and comments (tools/check_artifacts.py, in every release bundle);
  • the exit criteria of stage 3 of the playbook Project foundation are met;
  • the owner has approved it, where the playbook asks for approval. Only the owner sets approved.

Source

templates/prd.md, as in the repository. Placeholders are in square brackets; comments carry the instructions.

prd.md
---
template: prd/2
title: Product Requirements
status: draft # draft | review | approved — only the owner sets approved
version: 0.1.0
date: YYYY-MM-DD
owner: Product Owner
tier: standard # lightweight | standard | high-assurance — the owner decides (mgmt.governance-tiers); the reason is in the Blueprint
---
# PRD — Product name
<!-- Template from the UASK Engineering Hub, version 2, playbook mgmt.project-foundation, stage 3.
Save as project/PRD.md. The front matter says which tier the project is on. A section opening with
"Required: <tiers>" is required at those tiers only; "Optional" means never required; no marker means required
at every tier. Delete these comments when a section is written. The earlier version is
templates/prd-v1.md; a PRD without `template:` is checked against it and gets a warning. -->
Keywords MUST, SHOULD, MAY follow [RFC 2119](https://www.rfc-editor.org/rfc/rfc2119) /
[RFC 8174](https://www.rfc-editor.org/rfc/rfc8174). Ids (`S-…`, `FR-n`, `NFR-n`, `AC-n`; outcomes are `O-n` in
the vision) are stable, never reused, and referenced from the blueprint, roadmap, issues and tests.
## Problem
<!-- Who has which problem today, and what it costs them. Two to five sentences, no solution yet. -->
## Users and context
<!-- Optional. User groups, their environment and devices, existing systems the product lives next to. -->
## Scenarios
<!-- End-to-end situations the product must handle; each serves at least one vision outcome, named in the
last column. Every outcome of the vision must appear there at least once. -->
| ID | Scenario | Actor | Inputs | Result | Outcome |
|---|---|---|---|---|---|
| S-A | … | … | … | … | O-1 |
## Functional requirements
<!-- One testable statement per row; cite the scenario. Priority is must, should or could (what is dropped first
when the appetite runs out). State is for a product that is already partly or fully built (done, partly,
not started, waiting), so that the table also shows what is left; for a new product write "not started". -->
| ID | Requirement | Scenario | Priority | State |
|---|---|---|---|---|
| FR-1 | The system MUST … | S-A | must | not started |
## Non-functional requirements
<!-- Quality requirements with measurable targets. Name the characteristic in the nine words of ISO/IEC 25010:2023
(item arch.quality-attributes#R1): functional suitability, performance efficiency, compatibility, interaction
capability (accessibility belongs here), reliability, security, maintainability, flexibility, safety.
"Fast" is not a requirement; "p95 page load under 2 s on <device>" is. Say how each is checked (a test, a
monitor, an inspection or a review). Recovery, rollback, monitoring and support dates go in "Operability"
below. -->
| ID | Characteristic | Requirement | Checked by |
|---|---|---|---|
| NFR-1 | … | … | … |
## Acceptance criteria
<!-- Checks that decide whether the release is accepted; each names the ids it proves. Each also states how it
is performed: the method, the environment where it can really be performed, the expected result (a number
or an observable state, not a distribution) and who performs it. Check that it can be performed in that
environment: a criterion that needs real personal data on a test site, or two outcomes that cannot happen
together, is a defect of the criterion. Example:
- [ ] AC-3 Card payment completes (proves FR-7). Method: test card, sandbox of the payment service;
environment: test site; expected: order status "paid" within 60 s; performed by: owner. -->
- [ ] AC-1 …
## Constraints and assumptions
<!-- Optional. Fixed givens (platform, hosting, budget, deadline) and unverified beliefs; each assumption
gets a planned check. -->
## Parity with the current system
<!-- Optional. For a re-foundation of a running product: one table of what already works instead of dozens of
requirements. Write requirements only for what changes, what is at risk and how the switch is made. -->
| Capability | State | Evidence |
|---|---|---|
| … | done / partly / not started | … |
## External interfaces and environment
<!-- Required: standard, high-assurance — and at any tier where the work replaces, migrates or rebuilds a running
system (playbook mgmt.project-foundation, R12). List everything outside the system that depends on it or on
its hosting, and end each entry as a requirement or acceptance criterion, or as a decision to drop it. Walk the
list: public addresses and redirects (including with and without "www", http and https, rewrite files); feeds
and sitemaps consumed by others; webhooks and notification addresses of payment, delivery and other
services; consumers of any API; scheduled jobs; settings kept in the hosting panel (language version and
handler, limits, mail routing); DNS records and certificates; outgoing mail and the services that depend
on it; copies and backups kept outside the host; accounts and keys held by third parties. The Blueprint's
section 3 carries the same inventory with a contract or an owner for each. -->
| Interface or setting | Today | After the switch | Who owns it | Covered by |
|---|---|---|---|---|
| … | … | … | … | FR-n / AC-n / dropped |
## Operability
<!-- At every tier — what the people who run the system need, as values (item
arch.quality-attributes#R8, playbook mgmt.project-foundation, R13): a number, a date, a named owner, or "none"
with the reason. Derive each value from what the stakeholders can bear, not from what the hosting offers; where
a value is a guess, say so and name the check that settles it. At the lightweight tier, "none" with a reason is
enough for a point that does not apply, but a system that holds data the business cannot recreate still states the
recovery point, the recovery time and the restore test (playbook mgmt.project-foundation, R13). The Blueprint's
section 7 says how each value is met. -->
| Point | Requirement (value) | Checked by |
|---|---|---|
| Availability | the target and the hours it applies to | … |
| Recovery point (RPO) | how much data may be lost | … |
| Recovery time (RTO) | how long until it is back | … |
| Restore test | how often, by whom, and the date of the last proof | … |
| Rollback | how a release is undone and how long that takes | … |
| Monitoring | what is watched and who is told | … |
| Who runs it | the person or team, and the cover | … |
| End of life | the end-of-support date of each runtime and platform, and the plan | … |
## Sources and traceability
<!-- Optional. Where each requirement came from: the client's brief, a numbered clause of a specification, an
owner decision with its date. Needed when requirements are derived from client material or earlier
decisions rather than invented. -->
| Requirement | Source |
|---|---|
| FR-n | … |
## Out of scope
<!-- What this release will not do, following the vision's non-goals. -->
- …
## Open questions
<!-- Optional. Each question names who answers it (owner, client through the owner, technical lead, an expert of
the existing system) and by when. -->