Prompts

Phase 1: Initial Brainstorm

I want to build [your idea]. Here's everything I have so far:

[Describe your vision, users, key features, constraints]

I'd like to have a proper brainstorming conversation to:
1. Define the MVP scope (what's essential vs nice-to-have)
2. Decide on tech stack with rationale
3. Sketch the data architecture
4. Identify the biggest technical risks

Please start by asking me the questions you need answered
to do this properly. Don't hold back.

Phase 2: Generate Foundation Documents

Based on everything we've discussed, please generate all
the foundation documents for this project:

1. README.md (overview, setup, current phase)
2. ARCHITECTURE.md (full schemas, routes, decisions)
3. LEARNINGS.md (empty template)
4. .clinerules (iron-clad quality rules)
5. SPRINT_RULES.md
6. TASK_TEMPLATE.md
7. Sprint 1 plan (task index with estimates)
8. Individual task specs for each Sprint 1 task

For ARCHITECTURE.md: include actual database schemas with
all columns, types, and constraints. Not descriptions.

For .clinerules: mandatory testing (not "try to test"),
plan mode always, confidence scoring, no scope creep.
Make the rules strict enough that AI won't skip them.

Phase 3: Execute a Task

Can we please plan task [X.X] from the sprint plan?

[Optional: any additional context not in the docs, e.g.
external API details, changed requirements, specific constraints]

Phase 3: Fix or Debug (Critical)

I need to fix this issue: [describe the problem clearly]

Current behaviour: [what happens]
Expected behaviour: [what should happen]
Error message (if any): [exact error]

Please investigate the relevant code and propose a plan
before making any changes. Do not start fixing until we've
agreed on the approach.

Phase 3: New Feature (Mid-Project)

I want to add [feature] to the project. The GitHub repo
is synced in this Project so you have full context.

Please:
1. Review how this feature fits with the current architecture
2. Identify what needs to change
3. Propose the implementation approach
4. Generate task specs for the sprint plan

Ask any questions you need before proposing the approach.

Phase 4: Phase Audit

Please do a structured code review of this project.
You are a senior reviewer, NOT the developer who built it.
Be direct and critical. Do not soften findings.

Review for:
1. Security vulnerabilities
2. Error handling gaps
3. Performance concerns
4. Code consistency
5. Missing tests
6. Documentation accuracy vs codebase
7. Anything that would concern a senior engineer

Output: prioritised list of findings (Critical/High/Medium/Low)
with specific file references where possible.

Context Rescue

We're going in circles on this. Before we start a fresh
conversation, please write a context rescue document:

1. What we were trying to accomplish
2. What approaches we've tried and why they didn't work
3. Current state of the relevant code
4. What the next debugging steps should be
5. Any relevant discoveries or constraints

Format it as a task doc I can use to start a fresh conversation.

Documentation Update

Please review our documentation vs the current codebase:

1. Is ARCHITECTURE.md accurate? What's changed?
2. Is .clinerules still relevant?
3. What should be added to LEARNINGS.md based on this sprint?
4. Are task specs for completed tasks still useful to keep?

Propose specific updates for each document.

Sprint Planning for New Sprint

Sprint 1 is complete. The repo is synced so you can see
what was built.

Please help plan Sprint 2:
1. Review what Sprint 1 built and note any gaps
2. Consider the overall project roadmap
3. Propose Sprint 2 tasks in priority order
4. Flag any dependencies between tasks
5. Generate task specs once we agree on the sprint plan