Generalizability audit — what project types does Klonode fail on?

[smorchj]

Klonode was built and tested against one project type (Next.js + Prisma + Three.js + Postgres). This is the tracker for "what does Klonode produce empty or useless CONTEXT.md for today, and how do we close the gap." We need help from people who actually use these stacks day to day.

How to ship a fix (it's small)

The pattern is now well-trodden. Adding a language to the content extractor is one regex block in one file. PR #13 added Python, Java, and Ruby in 104 lines total — read that PR end-to-end and you'll have the template:

• Reference PR: Add Python, Java, and Ruby export extraction #13 — adds 3 languages to packages/core/src/analyzer/content-extractor.ts
• Reference tests: packages/core/src/analyzer/content-extractor.test.ts (look for the describe('Python') / describe('Java') / describe('Ruby') blocks)
• Quality bar: see CONTRIBUTING.md

Then for whichever language or framework you pick:

1. Hand-write a minimal source fixture (one file, 10–30 lines)
2. Add a regex block to extractExports (or extractApiRoutes, detectPatterns, etc. depending on what you're matching)
3. Add tests asserting the exports come out the way you expect
4. pnpm --filter @klonode/core test — must pass
5. Open a PR linking the gap issue (Closes #17 etc.)

That's it. No AST parser dependency, no new packages. Match the style of the file you're editing.

Already shipped (so you can see the velocity)

Language / framework	PR	LOC
Python, Java, Ruby extractors	#13	+104
Astro, Deno, Bun detection	#12	–
Prisma schema parser	(in #12 / #13 era)	–

Gaps with their own sub-issue (good first contributions, ordered by demand)

• Go support — Add Go support to content extractor #17 (high demand, simple pattern)
• Rust support — Add Rust support to content extractor #18 (high demand)
• PHP support — Add PHP support to content extractor #15
• C# support — Add C# support to content extractor #16
• Kotlin support — Add Kotlin support to content extractor #19
• Detect Rails in tool-detector — Detect Rails in tool-detector #25
• Detect Django in tool-detector — Detect Django in tool-detector #24
• Detect FastAPI in tool-detector — Detect FastAPI in tool-detector #23
• Detect Remix in tool-detector — Detect Remix framework in tool-detector #20
• Detect SolidJS / SolidStart in tool-detector — Detect SolidJS / SolidStart in tool-detector #21
• Detect Qwik / QwikCity in tool-detector — Detect Qwik / QwikCity in tool-detector #22
• YAML config parsing — Add YAML config file parsing #26 (GitHub Actions workflows, Compose, k8s)
• TOML config parsing — Add TOML config file parsing (Cargo.toml, pyproject.toml) #27 (Cargo.toml, pyproject.toml — extract deps, features, workspace)

Gaps without a dedicated issue yet

These are real gaps but no one has filed a sub-issue. If you want to claim one, comment here first so we don't double up, then open a focused sub-issue.

Languages with extension detection but no extractor

• Swift (.swift), Dart (.dart), Scala, Clojure, Elixir, Haskell, F#, Crystal, Zig

Schema sources we don't parse

• Drizzle ORM (schema.ts DSL) — we detect the framework but don't read the definitions
• TypeORM entities, Sequelize models, Mongoose schemas (TS/JS)
• SQLAlchemy models (Python), Django models, Rails ActiveRecord models
• GORM (Go), Diesel (Rust), Ecto (Elixir)

Infrastructure-as-code

• Terraform .tf, Pulumi
• Kubernetes YAML manifests, Helm charts
• CloudFormation, Ansible playbooks
• Dockerfile content (we detect existence but not the steps)
• Caddyfile, nginx.conf, systemd units

Other config formats

• OpenAPI / Swagger, AsyncAPI, JSON Schema, Protocol Buffers
• package.json scripts field (we only use deps today)

Build / tooling configs

• Makefile targets, Justfile
• Gradle (build.gradle), Maven (pom.xml)
• Rake (Rakefile)

Mobile / desktop / game

• Unity .unity / .meta, Unreal .uproject, Godot .tscn
• Shaders (.glsl, .wgsl, .hlsl)
• Flutter project structure
• Android manifests, iOS Info.plist

Documentation as structure

• Markdown / MDX — headings as navigation structure
• Jupyter notebooks (.ipynb) — code cells + markdown cells

Monorepo tooling

• Nx, Lerna, Rush, Bazel / Buck / Pants

Done condition

Close this once Rails, Django, Go backends, Rust CLIs, Flutter apps, and Terraform monorepos all produce useful CONTEXT.md.

---

Currently most needed: Go (#17), Rust (#18), Django detector (#24). If you use any of those daily and want a small, well-scoped first contribution to an active OSS project, this is the one.

[smorchj]

Re-opening this for help.

If you use any of these stacks day to day, this is the easiest possible PR to land in an active OSS project right now:

• Go backends → Add Go support to content extractor #17
• Rust CLIs → Add Rust support to content extractor #18
• Django apps → Detect Django in tool-detector #24
• Rails apps → Detect Rails in tool-detector #25
• PHP projects → Add PHP support to content extractor #15
• C# / Kotlin → Add C# support to content extractor #16, Add Kotlin support to content extractor #19
• FastAPI → Detect FastAPI in tool-detector #23

The pattern is one regex block in one file. Reference PR #13 added Python, Java, and Ruby in 104 lines total — read it end-to-end and you'll have the template. Tests are right there in content-extractor.test.ts. We do not auto-decline first-time contributors and we'll leave actionable review feedback rather than closing — see the new How PRs are reviewed section.

Pick whichever sub-issue matches a stack you actually use, comment on it to claim it, open a PR.