Project Discovery

Project Discovery

Build a practical understanding of a project before proposing or making changes.

Workflow

1. Identify the goal

   • Determine whether the user needs an overview, architecture summary, stack analysis, risk assessment, implementation plan, migration direction, or onboarding notes.

2. Inspect high-signal files

   • Read README.md, project instructions, package manifests, lockfiles, docs, architecture notes, CI config, test config, and main entry points.
   • Prefer actual files over guesses — the research-methodology source-tier idea applies here too: a file in the repo outranks a README's description of what the repo does, which outranks an outside blog post about the project.

3. Map the system

   • Identify core modules, data flow, user-facing surfaces, external services, build commands, test commands, deployment hints, and local conventions.
   • Note areas that need follow-up inspection instead of pretending certainty.

4. Assess readiness

   • Call out missing docs, unclear ownership, brittle tests, risky dependencies, weak validation, or hidden configuration.
   • Keep recommendations incremental.

5. Report

   • Respond in the user's language.
   • Link to files when possible.
   • End with a practical next step.

Output Format

# Project Discovery

## What This Project Does

## Stack

## Important Files

## Architecture Notes

## Development Workflow

## Risks And Unknowns

## Recommended Next Steps

Rules

• Do not propose broad rewrites before understanding local patterns.
• Distinguish verified facts from assumptions.
• Keep the report useful for the next action, not encyclopedic.