smithery.ai

TalentFilter Developer

Expert knowledge required to develop features for the TalentFilter SaaS platform, including architecture, security, and tech stack compliance.

Installation

$ npx skills add https://smithery.ai

Similar popular skills

Related neighbors and high-traction skills in the same topics — useful to compare before installing.

Also in this package

Other skills from smithery.ai · top by installs.

npx skills add https://smithery.ai

Browse all from smithery.ai

More details

Agent compatibility

Declared targets from SKILL.md / docs. Unmarked agents are not listed — the skill may still install via the CLI.

Claude Code Not declared
Cursor Not declared
Codex Not declared
GitHub Copilot Not declared
Windsurf Not declared
Gemini CLI Declared
Cline Not declared
OpenCode Not declared

Skill metadata

Parsed from SKILL.md frontmatter.

Declared agents gemini

Package contents

Files included with this skill beyond the listing page.

  • skill md SKILL.md 4,328 B
  • docs SUMMARY.md 172 B

History

  1. First recorded snapshot · 1 installs

SKILL.md

TalentFilter Development Skill

Use this skill whenever you are generating code, planning features, or debugging issues for the TalentFilter project. It contains the authoritative architectural decisions and patterns.

1. Technology Stack Compliance

Frontend (User & Recruiter)

  • Framework: Next.js 16.1.4+ (App Router). CRITICAL: Versions below 16.1.4 are vulnerable to RCE and Source Code Exposure (CVE-2025-66478, CVE-2025-55183). Always verify against the latest security advisories.
  • Styling: Tailwind CSS + Shadcn/UI (Radix primitives). Use clsx and tailwind-merge for class manipulation.
  • State Management:

- Auth: Primary session stored in HTTP-Only Cookies (tf_session). - Zustand: Sync client-side ephemeral state (Auth metadata, UI toggles). - TanStack Query: Use for ALL server data fetching. No distinct useEffect data fetching.

  • Internationalization: next-intl for EN/DE support.
  • Animations: framer-motion for premium micro-interactions.
  • Icons: Lucide React.

Backend (API & AI)

  • Framework: FastAPI (Python 3.10+). Fully async.
  • AI Integration: Google Gemini API via google-generativeai.

- Safety: NEVER call AI from the frontend. Always proxy via Backend AIService. - Reliability: Use JSON Mode for structured outputs.

  • Validation: Pydantic V2 for all Request/Response models.

Database & Auth

  • Service: Supabase.
  • Auth: Supabase Auth (handled via client/headers).
  • Security: Row Level Security (RLS) is mandatory.

- Recruiters: auth.uid() = recruiter_id - Candidates: NO direct DB access. All writes via Backend Service Role.

2. Architectural Patterns

Frontend: Locale-Based Feature Structure

Organize code by business feature within the internationalization wrapper.

  • Routing: Locale-based routing (src/app/[locale]).
  • Redirection: Use Server-Side Layouts (layout.tsx) with cookies() for authoritative auth checks and redirects. Use proxy.ts (Next.js 16) only for lightweight routing.
  • State: Zustand for client state, TanStack Query for server state.
  • i18n: All UI text must be in messages/*.json. Use @/i18n/routing for navigation.
  • Landing Page: Must use high-impact visual design (Gradients, Motion) and clear CTAs.
src/components/features/
  interview/       # Timer, Recorder, QuestionCard
  builder/         # JobForm, PromptEditor

Backend: Layered Architecture

Strictly separate concerns.

  1. Routers (app/api/): Parse input, call service, return response.
  2. Services (app/services/): Business logic, AI calls, DB transactions.
  3. Schemas (app/schemas/): Pydantic models.
  4. Models (app/models/): Database/ORM definitions.

3. Security & Anti-Cheating Implementation

CRITICAL: This is a competitive screening tool.

  1. Visibility Tracking: Use the visibilitychange API as detailed in the security skill.
  2. Input Lockdown: Prevent Copy/Paste on answer fields via the usePreventCopyPaste hook.
  3. Timers: Trust the Backend timestamp. The frontend timer is just for UI.
  4. Request Limiting: All AI-heavy endpoints must implement rate limiting to control costs and prevent abuse.
  5. Skill Reference: Always refer to .agent/skills/security/SKILL.md for the full implementation details of these measures.

4. Development Workflow

  1. Plan: Check ImplementationPlan.md for current phase.
  2. Simulate: Before writing complex logic, briefly outline the component/service structure.
  3. Implement: Write clean, typed code (TypeScript/Python).
  4. Verify: Ensure RLS policies cover new tables.

5. Important References

  • docs/FRONTEND_ARCHITECTURE.md: Detailed state management patterns.
  • docs/BACKEND_ARCHITECTURE.md: API design and AI Engine specs.
  • docs/UNIFIEDWIZARDSTRATEGY.md: The new AI-first, human-refined Job Creation approach.
  • ImplementationPlan.md: The roadmap.
  • .agent/workflows/unified-wizard-implementation.md: Step-by-step implementation guide for the Unified Wizard.