AI-native scaffolding: grit describe "<feature>" → propose models, handlers, UI

Open
#53 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Assessment

Difficulty
5/5
Estimated time
Over a week
Newbie friendliness
10/100
Issue type
Feature
Clarity
Needs clarification
Activity status
Stale
Tech stack
go
Domain
ai, cli, tooling

Research direction

Start by reading the existing grit CLI entry point and the grit generate implementation to see how generators, schema, permissions and AGENTS.md conventions are wired. The issue describes a new grit describe command that calls Claude and produces a plan, but it names no files or tests, so first map the command dispatch, generator and permission catalogue. Done means a reviewed implementation plan for the command, not generated code.

Written by the indexing model from the issue text.

Description

enhancement help wanted

Vision

This is the killer feature. The next generation of frameworks (v0, Convex, base44) blur the line between "describe" and "build". Grit has a unique advantage: the codebase has strong conventions (TwoPane, drawer forms, RBAC, generators) that an AI can reliably extend.

Proposal

```
grit describe "tenants need a vacation hold feature — they pause their booking
for N days without paying, can extend by 7 days max, requires manager approval"
```

The CLI:

  1. Reads the existing schema + permission catalogue + AGENTS.md conventions
  2. Calls Claude with a structured prompt + the relevant codebase slice
  3. Proposes a plan: new models, new endpoints, new permissions, new UI sections, new tests
  4. User reviews + accepts → `grit generate` is invoked with the parsed plan
  5. Output is committed to a `grit/` branch with a generated PR description

Critically: the framework's strict conventions (issues #18-#36) make AI output high-quality. Most AI scaffolders produce bespoke code; Grit's would produce code indistinguishable from hand-written framework code because the framework constrains it.

Why this is differentiating

Encore and Convex talk about AI-native infra. Nobody has nailed it for self-hosted Go. "Describe a feature in English, get a working PR" is the kind of demo that goes viral.

Acceptance

  • The vacation-hold example produces compiling, type-checked code with appropriate permission gates
  • Generated code passes the framework's lint rules (issue #5 / #18)
  • A senior engineer reviewing the PR sees "this looks like our team wrote it"
Dominant language
Go
Stars
152
Forks
20
Avg merge
1d 49m
Merged PRs (30d)
5

Getting set up

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

More from MUKE-coder/grit

All issues in MUKE-coder/grit

Similar issues

More Go issues

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.