By kulykivska
Per-stack reviewer subagents: Python/FastAPI, web (Vite/React/TS), SwiftUI, ML pipeline.
Reviews ML pipeline diffs (feature engineering, training, calibration, prediction serving) for leakage, regression-gate compliance, and cross-repo contract drift. Use for any change touching models, features, or training code. Returns concrete file:line findings.
Reviews Python/FastAPI diffs (RaceModel API, Jarvis, predictors) for bugs, conventions, and failure handling. Use to review backend Python changes before commit/push. Returns concrete file:line findings.
Reviews SwiftUI/iOS diffs (RaceModel iOS, Aura, GT7 Engineer) for crashes, modern-API usage, and StoreKit/paywall correctness. Use to review Swift changes before commit/push. Returns concrete file:line findings.
Reviews web frontend diffs (Vite/React/TypeScript SPAs) for bugs, UX/policy conventions, and responsiveness. Use to review frontend changes before commit/push. Returns concrete file:line findings.
One source of truth for personal Claude Code tooling across all projects
(RaceModel, f1-predictor, Jarvis, Aura, GT7 Engineer, ...). Registered as the
personal marketplace from this local path and enabled by default in
~/.claude/settings.json, so every project picks it up automatically.
| Plugin | Type(s) | What it does |
|---|---|---|
pre-push-review | skill | Fix-and-verify review before any git push (pairs with the user-level pre-push gate hook). |
morning | skill | Daily kickoff recap from git + transcripts + memory. |
guardrails | blocking hooks | Vetoes destructive Fly ops (destroy/scale-0/secrets unset), DROP/TRUNCATE via DB clients, force-push to main, AI attribution in commits, secrets/PII in staged + pushed diffs. |
coach | non-blocking hooks | Per-stack nudges on edit (Python/TS/Swift, FEATURE_COLS sync reminder, no long dashes in UI text), session banner, failure hints (ports, Docker reload, Fly release_command), uncommitted reminder. |
sdlc | skills + subagents | requirements → plan-task → implement → qa → task-review loop, plus architect and debugger subagents. |
reviewers | subagents | python-reviewer, web-reviewer, swiftui-reviewer, ml-reviewer (LORO gate, leakage, cross-repo FEATURE_COLS sync). |
fly-ops | skills | deploy (incl. release_command gotcha), fly-logs, incident triage with known failure modes. |
lsp | LSP | pyright, typescript-language-server, sourcekit-lsp (see SETUP.md for binaries). |
monitors | monitors | tsc-watch build errors + .dev.log error tail, idle-safe. |
mcp-catalog | MCP example | GitHub / Playwright / Postgres catalog; copy into a project's .mcp.json to activate. |
~/.claude/settings.json:
{
"extraKnownMarketplaces": {
"personal": { "source": { "source": "directory", "path": "<this repo's absolute path>" } }
},
"enabledPlugins": {
"pre-push-review@personal": true,
"morning@personal": true,
"...": true
}
}
Note: the git push gate itself stays as a user-level PreToolUse hook
(~/.claude/scripts/pre-push-gate.sh), NOT in the guardrails plugin, to avoid
double-gating in repos that register an equivalent gate of their own.
.claude/).claude/ in this repo is the versioned copy of the machine's user-level
Claude Code config:
.claude/settings.json: copy of ~/.claude/settings.json.claude/scripts/pre-push-gate.sh: the push gate, generic (every project,
every branch, no per-project exemptions); the live copy is
~/.claude/scripts/pre-push-gate.sh.claude/scripts/statusline.sh: status line scriptThe repo copy is the source of truth: edit here, then sync to ~/.claude
(cp .claude/scripts/* ~/.claude/scripts/). On a new machine: copy
.claude/settings.json + .claude/scripts/ into ~/.claude/ and adjust the
absolute marketplace path.
python3 scripts/validate.py
Edit here, bump the plugin's version, restart Claude Code (local directory
marketplace: changes are picked up on reload). New plugin: folder +
.claude-plugin/plugin.json, register in .claude-plugin/marketplace.json,
enable in ~/.claude/settings.json.
Uses power tools
Uses Bash, Write, or Edit tools
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
npx claudepluginhub kulykivska/claude-plugins --plugin reviewersNon-blocking coaching hooks: per-stack nudges on edit, session status banner, failure hints, uncommitted-changes reminder.
Fly.io ops runbooks: deploy, fly-logs, incident triage.
Business growth subagents: aso-optimizer (App Store), growth-analyst (funnel/conversion), outreach-writer (partnership and consulting pitches).
Language servers: Python (pyright), TypeScript, Swift. Binaries installed per machine, see SETUP.md.
Fix-and-verify review that fixes code smells, security, scalability, and performance issues before any git push.
Ultra-compressed communication mode. Cuts 65% of output tokens (measured) while keeping full technical accuracy by speaking like a caveman.
Comprehensive UI/UX design plugin for mobile (iOS, Android, React Native) and web applications with design systems, accessibility, and modern patterns
Agent that simplifies and refines code for clarity, consistency, and maintainability while preserving functionality
Multi-model consensus engine integrating OpenAI Codex CLI, Gemini CLI, and Claude CLI for collaborative code review and problem-solving.
Unified capability management center for Skills, Agents, and Commands.