Language-agnostic AI knowledge registry for Technical Project Management, PRDs, PRD review, Software Architecture planning, Task breakdown, Estimation, Risk assessment, Status reporting, Backlog prioritization, Sprint planning, Retrospectives, and Agile ticket generation. Uses Markdown + Front-matter architecture.
95
97%
Does it follow best practices?
Impact
95%
1.03xAverage score across 10 eval scenarios
Passed
No known issues
Short discovery shape (when the user wants a one-pager before a full PRD):
Name: Search Autocomplete Summary: Improve search UX by providing real-time autocomplete suggestions.
Problem: Users struggle to find items quickly; searches without suggestions return many irrelevant results.
Goal: Increase successful search completions by 15% in 3 months.
User story: As a user, I want inline suggestions while typing so I can select a result without completing the full query.
Acceptance criteria:
Rollout: Feature flag, 10% → 50% → 100% with metrics check at each step.
Checklist:
Input (feature description): Add the ability for admins to export user data as a CSV file from the admin dashboard.
Output:
# PRD: Admin CSV User Export
## Introduction
Admin users need to export a filtered list of user records as CSV from the admin dashboard for audits and reporting, without direct database access.
## Goals
- Admins can download user data without direct database access.
- Export completes in under 5 seconds for up to 10,000 records.
- Exported data respects existing admin permission scopes.
## User Stories
- As an admin, I want to export all users as a CSV so that I can share data with the ops team.
- As an admin, I want to filter users before exporting so that I only download relevant records.
## Functional Requirements
1. The system must show an **Export CSV** action on the admin users index page.
2. The system must trigger a `.csv` download when the admin confirms export, including columns: id, email, created_at, role, status.
3. The system must apply the same filters as the index (e.g. role, status) to the exported row set.
## Non-Functional Requirements
- Only users with the `admin` role can trigger an export.
- Exports larger than 10,000 rows must run via a background job and notify the admin (e.g. email with download link) to avoid request timeouts.
## Non-Goals
- Exporting associated records (orders, sessions, etc.).
- Custom column selection in v1.
- Scheduled or automated exports.
## Design Considerations
- Reuse existing admin index filters and table affordances; export entry point should be obvious and hard to trigger accidentally.
## Technical Considerations
- Use background job processing for large exports; consider streaming CSV for medium sizes if product wants synchronous download for <10k rows only.
## Implementation Surface
Areas likely touched (no code — discovery only):
- Backend — admin user controller or equivalent endpoint
- Data layer — `User` scopes for filtering
- Services — optional `UserCsvExport` orchestration
- Background jobs — export + delivery when row count exceeds threshold
- Notifications — export ready / link to file
- External integrations — none
## Success Metrics
- Median export time < 5s for ≤10k visible rows on staging hardware.
- Zero unauthorized export attempts succeeding (covered by policy/request specs).
## Open Questions
- Should each export be logged for audit (who, when, filter snapshot)?
- Hard cap on rows per export (product/legal)?
## Next Steps
Recommended next step: **generate-tasks** — break this PRD into implementation tasks with TDD gates.
Alternative chain targets: `plan-tickets`.docs
evals
scenario-1
scenario-2
scenario-3
scenario-4
scenario-5
scenario-6
scenario-7
scenario-8
scenario-9
scenario-10
skills
backlog
prioritize-backlog
ceremony
create-retrospective
plan-sprint
prd
create-prd
review-prd
task-management
estimate-tasks
generate-tasks
plan-tickets