senior-fullstack
Fullstack development toolkit with project scaffolding for Next.js, FastAPI, MERN, and Django stacks, code quality analysis with security and complexity scoring, and stack selection guidance. Use when the user asks to "scaffold a new project", "create a Next.js app", "set up FastAPI with React", "analyze code quality", "audit my codebase", "what stack should I use", "generate project boilerplate", or mentions fullstack development, project setup, or tech stack comparison.
- License 许可
- MIT
- 最近更新
- 2026-07-14
- Stars
- ★ 0
- Forks
- 0
01 前言 Front Matter
- name
- senior-fullstack
- description
- Fullstack development toolkit with project scaffolding for Next.js, FastAPI, MERN, and Django stacks, code quality analysis with security and complexity scoring, and stack selection guidance. Use when the user asks to "scaffold a new project", "create a Next.js app", "set up FastAPI with React", "analyze code quality", "audit my codebase", "what stack should I use", "generate project boilerplate", or mentions fullstack development, project setup, or tech stack comparison.
02 资源文件 12
- 参考文档profiles/enterprise-scale.json3.5 KB
- 参考文档profiles/internal-tool.json2.5 KB
- 参考文档profiles/marketing-site.json2.8 KB
- 参考文档profiles/saas-startup.json2.7 KB
- 参考文档references/architecture_patterns.md12.7 KB
- 参考文档references/composition_map.md5.0 KB
- 参考文档references/development_workflows.md16.4 KB
- 参考文档references/forcing_questions.md7.9 KB
- 参考文档references/tech_stack_guide.md13.8 KB
- 脚本scripts/code_quality_analyzer.py21.9 KB
- 脚本scripts/fullstack_decision_engine.py14.8 KB
- 脚本scripts/project_scaffolder.py26.7 KB
03 SKILL.md 正文
Senior Fullstack
Fullstack development skill with project scaffolding and code quality analysis tools.
Table of Contents
Trigger Phrases
Use this skill when you hear:
- "scaffold a new project"
- "create a Next.js app"
- "set up FastAPI with React"
- "analyze code quality"
- "check for security issues in codebase"
- "what stack should I use"
- "set up a fullstack project"
- "generate project boilerplate"
Tools
Decision Engine
Deterministic profile picker. Given four assumptions (team-size, cadence, user-facing, budget) plus optional traffic/sensitivity inputs, ranks the four built-in profiles and returns the matched profile with SLO floor and named approver chain. Refuses to recommend a profile without the four required inputs.
Usage:
# See all options
python scripts/fullstack_decision_engine.py --help
# Run against a sample input
python scripts/fullstack_decision_engine.py --sample
# Pick a profile from real inputs
python scripts/fullstack_decision_engine.py \
--team-size-12mo 8 --cadence daily --user-facing true --budget 5000 \
--traffic-p99-rps 50 --data-sensitivity pii-only
# JSON output for downstream tools
python scripts/fullstack_decision_engine.py --sample --output json
Returns: matched profile name, score, matched/violated constraints, stack recommendation, anti-recommendations, SLO floor, named-approver chain, and canon references.
The engine encodes the same matrix the conversational grill walks through — use it directly when inputs are already known, or via the cs-fullstack-engineer agent for the question-by-question grill.
Project Scaffolder
Generates fullstack project structures with boilerplate code.
Supported Templates:
nextjs- Next.js 14+ with App Router, TypeScript, Tailwind CSSfastapi-react- FastAPI backend + React frontend + PostgreSQLmern- MongoDB, Express, React, Node.js with TypeScriptdjango-react- Django REST Framework + React frontend
Usage:
# List available templates
python scripts/project_scaffolder.py --list-templates
# Create Next.js project
python scripts/project_scaffolder.py nextjs my-app
# Create FastAPI + React project
python scripts/project_scaffolder.py fastapi-react my-api
# Create MERN stack project
python scripts/project_scaffolder.py mern my-project
# Create Django + React project
python scripts/project_scaffolder.py django-react my-app
# Specify output directory
python scripts/project_scaffolder.py nextjs my-app --output ./projects
# JSON output
python scripts/project_scaffolder.py nextjs my-app --json
Parameters:
| Parameter | Description |
|---|---|
template |
Template name (nextjs, fastapi-react, mern, django-react) |
project_name |
Name for the new project directory |
--output, -o |
Output directory (default: current directory) |
--list-templates, -l |
List all available templates |
--json |
Output in JSON format |
Output includes:
- Project structure with all necessary files
- Package configurations (package.json, requirements.txt)
- TypeScript configuration
- Docker and docker-compose setup
- Environment file templates
- Next steps for running the project
Code Quality Analyzer
Analyzes fullstack codebases for quality issues.
Analysis Categories:
- Security vulnerabilities (hardcoded secrets, injection risks)
- Code complexity metrics (cyclomatic complexity, nesting depth)
- Dependency health (outdated packages, known CVEs)
- Test coverage estimation
- Documentation quality
Usage:
# Analyze current directory
python scripts/code_quality_analyzer.py .
# Analyze specific project
python scripts/code_quality_analyzer.py /path/to/project
# Verbose output with detailed findings
python scripts/code_quality_analyzer.py . --verbose
# JSON output
python scripts/code_quality_analyzer.py . --json
# Save report to file
python scripts/code_quality_analyzer.py . --output report.json
Parameters:
| Parameter | Description |
|---|---|
project_path |
Path to project directory (default: current directory) |
--verbose, -v |
Show detailed findings |
--json |
Output in JSON format |
--output, -o |
Write report to file |
Output includes:
- Overall score (0-100) with letter grade
- Security issues by severity (critical, high, medium, low)
- High complexity files
- Vulnerable dependencies with CVE references
- Test coverage estimate
- Documentation completeness
- Prioritized recommendations
Sample Output:
============================================================
CODE QUALITY ANALYSIS REPORT
============================================================
Overall Score: 75/100 (Grade: C)
Files Analyzed: 45
Total Lines: 12,500
--- SECURITY ---
Critical: 1
High: 2
Medium: 5
--- COMPLEXITY ---
Average Complexity: 8.5
High Complexity Files: 3
--- RECOMMENDATIONS ---
1. [P0] SECURITY
Issue: Potential hardcoded secret detected
Action: Remove or secure sensitive data at line 42
Workflows
Workflow 1: Start New Project
- Choose appropriate stack based on requirements (see Stack Decision Matrix)
- Scaffold project structure
- Verify scaffold: confirm
package.json(orrequirements.txt) exists - Run initial quality check — address any P0 issues before proceeding
- Set up development environment
# 1. Scaffold project
python scripts/project_scaffolder.py nextjs my-saas-app
# 2. Verify scaffold succeeded
ls my-saas-app/package.json
# 3. Navigate and install
cd my-saas-app
npm install
# 4. Configure environment
cp .env.example .env.local
# 5. Run quality check
python ../scripts/code_quality_analyzer.py .
# 6. Start development
npm run dev
Workflow 2: Audit Existing Codebase
- Run code quality analysis
- Review security findings — fix all P0 (critical) issues immediately
- Re-run analyzer to confirm P0 issues are resolved
- Create tickets for P1/P2 issues
# 1. Full analysis
python scripts/code_quality_analyzer.py /path/to/project --verbose
# 2. Generate detailed report
python scripts/code_quality_analyzer.py /path/to/project --json --output audit.json
# 3. After fixing P0 issues, re-run to verify
python scripts/code_quality_analyzer.py /path/to/project --verbose
Workflow 3: Stack Selection
Use the tech stack guide to evaluate options:
- SEO Required? → Next.js with SSR
- API-heavy backend? → Separate FastAPI or NestJS
- Real-time features? → Add WebSocket layer
- Team expertise → Match stack to team skills
See references/tech_stack_guide.md for detailed comparison.
Reference Guides
Architecture Patterns (references/architecture_patterns.md)
- Frontend component architecture (Atomic Design, Container/Presentational)
- Backend patterns (Clean Architecture, Repository Pattern)
- API design (REST conventions, GraphQL schema design)
- Database patterns (connection pooling, transactions, read replicas)
- Caching strategies (cache-aside, HTTP cache headers)
- Authentication architecture (JWT + refresh tokens, sessions)
Development Workflows (references/development_workflows.md)
- Local development setup (Docker Compose, environment config)
- Git workflows (trunk-based, conventional commits)
- CI/CD pipelines (GitHub Actions examples)
- Testing strategies (unit, integration, E2E)
- Code review process (PR templates, checklists)
- Deployment strategies (blue-green, canary, feature flags)
- Monitoring and observability (logging, metrics, health checks)
Tech Stack Guide (references/tech_stack_guide.md)
- Frontend frameworks comparison (Next.js, React+Vite, Vue)
- Backend frameworks (Express, Fastify, NestJS, FastAPI, Django)
- Database selection (PostgreSQL, MongoDB, Redis)
- ORMs (Prisma, Drizzle, SQLAlchemy)
- Authentication solutions (Auth.js, Clerk, custom JWT)
- Deployment platforms (Vercel, Railway, AWS)
- Stack recommendations by use case (MVP, SaaS, Enterprise)
Quick Reference
Stack Decision Matrix
| Requirement | Recommendation |
|---|---|
| SEO-critical site | Next.js with SSR |
| Internal dashboard | React + Vite |
| API-first backend | FastAPI or Fastify |
| Enterprise scale | NestJS + PostgreSQL |
| Rapid prototype | Next.js API routes |
| Document-heavy data | MongoDB |
| Complex queries | PostgreSQL |
Common Issues
| Issue | Solution |
|---|---|
| N+1 queries | Use DataLoader or eager loading |
| Slow builds | Check bundle size, lazy load |
| Auth complexity | Use Auth.js or Clerk |
| Type errors | Enable strict mode in tsconfig |
| CORS issues | Configure middleware properly |
Assumptions and Verifiable Success Criteria (Karpathy discipline)
Before this skill scaffolds, recommends, or modifies any code, the following four assumptions MUST be surfaced. If any are unknown, the skill stops and walks the Forcing-question library instead.
- Team size today + 12-month headcount — drives architecture (monolith / modular / services). Sam Newman: "MonolithFirst."
- Deployment cadence target — drives CI/CD spend and feature-flag investment. Accelerate (Forsgren et al. 2018).
- User-facing vs. internal vs. marketing-site — drives stack pick and a11y/perf budget.
- Monthly cloud + SaaS budget ceiling — drives the build-vs-managed-service split.
Verifiable success criteria (Karpathy #4) — every recommendation this skill emits must include three machine-checkable numbers:
- An API latency target (p50, p95, p99 in ms)
- A frontend perf target (LCP, INP, CLS on mobile-4G)
- An uptime / SLO target
If any of those three is not stated, the recommendation is incomplete — go back to Q7 of the forcing-question library.
The scripts/fullstack_decision_engine.py tool encodes these checks: it refuses to recommend a profile without all four assumption inputs and prints the verifiable thresholds for the matched profile.
Customization profiles
Four built-in profiles in profiles/ calibrate every recommendation:
| Profile | When to pick | Cloud ceiling | Pattern |
|---|---|---|---|
saas-startup |
< 10 eng, customer-facing, daily+ cadence | $8K/mo | Modular monolith on Next.js + Postgres |
enterprise-scale |
50+ eng, regulated, per-PR with gates | $250K/mo | Domain-bounded services + platform team |
internal-tool |
≤ 5 eng, auth-walled, < 100 DAU | $500/mo | Retool-first; thin custom stack if forced |
marketing-site |
SEO-dependent, near-zero write | $200/mo | Static-first (Astro / 11ty / Next-static) |
Pick a profile via:
python scripts/fullstack_decision_engine.py \
--team-size 6 --team-size-12mo 12 \
--cadence daily --user-facing true --budget 5000 \
--traffic-p99-rps 45 --data-sensitivity pii-only
The tool returns the best-fit profile, the tradeoff against the runner-up (if within 15%), the stack recommendation, the anti-patterns to avoid on that profile, and the named-approver chain. This tool never auto-approves.
To add a custom profile: copy profiles/saas-startup.json to profiles/<your-org>.json, adjust the constraints and stack_recommendations blocks, and rerun. The JSON is the customization surface — no code changes needed.
Composition map
This skill does NOT reimplement scope owned by the POWERFUL-tier specialists. It forks into them. See references/composition_map.md for the full routing table. Key forks:
| Concern | Fork into |
|---|---|
| API contract review | engineering/skills/api-design-reviewer/ |
| Database schema design | engineering/skills/database-designer/ |
| Reliability / SLO design | engineering/slo-architect/ |
| CI/CD pipeline | engineering/skills/ci-cd-pipeline-builder/ |
| Performance profiling | engineering/skills/performance-profiler/ |
| Pre-commit Karpathy review | engineering/karpathy-coder/ |
| Pre-flight architecture grill | engineering/grill-me/ |
The cs-fullstack-engineer agent (in agents/engineering/cs-fullstack-engineer.md) orchestrates these forks via context: fork. Invoke it from another agent with Agent({subagent_type: "cs-fullstack-engineer", prompt: "..."}) or via the slash command /cs:fullstack-review <your problem>.
Forcing-question library (Matt Pocock grill)
Before locking any architecture or stack decision, walk the seven forcing questions in references/forcing_questions.md. Each has a recommended answer, canon citation, and kill criterion. The discipline:
- One question per turn. No bundling.
- Always recommend the answer with cited canon.
- Track answers in a working file (e.g.,
/tmp/fullstack-grill-<date>.md). - If a kill criterion trips, stop. Do not scaffold around an unresolved gap.
- After Q7, run
fullstack_decision_engine.pywith the seven answers as inputs.
Summary of the seven questions (full content in the reference):
- Team size today + 12-month headcount?
- Deployment cadence — per-PR, daily, weekly, quarterly?
- Customer-facing, internal tool, or marketing site?
- One-year p50 / p99 traffic forecast?
- Hiring against the stack or training the team?
- Year-one monthly cloud + SaaS ceiling?
- Three verifiable success criteria with numeric targets?
Invocation from other agents and skills
This skill is invokable by any other agent or skill via three surfaces:
- Slash command:
/cs:fullstack-review <prompt>— runs the full grill + decision engine + composition routing. - Agent subagent:
Agent({subagent_type: "cs-fullstack-engineer", prompt: "..."})— forks context, returns ≤ 200-word digest. - Direct tool call:
python scripts/fullstack_decision_engine.py ...— deterministic profile match without the conversational grill (use when inputs are already known).
See agents/engineering/cs-fullstack-engineer.md for the full invocation contract.
04 同仓库更多 Skills
syllabus
Generates a curated supplementary reading list from any course syllabus using Consensus academic search. Grill-me intake (syllabus input format + course audience + year range) plus a grouping forcing-options checkpoint before any search runs — so the reading list matches the course's level and recency need. Parses the syllabus to extract topics and learning outcomes, searches Consensus for recent peer-reviewed papers per topic, and produces a professionally formatted .docx with clickable Consensus links, plain-language summaries calibrated to audience level, and Bloom-higher-order discussion questions tied to course learning goals. Triggers whenever a user uploads a syllabus, course outline, or curriculum document and wants supplementary readings. Also triggers on: 'syllabus reading list', 'find papers for my course', 'create a reading list from this syllabus', 'recent research for my class', 'supplementary readings', 'find journal articles for these topics', 'what recent papers cover this material', 'any new research on these course topics', 'update my syllabus with recent papers'. Even casual mentions when a syllabus is attached should trigger this skill.
research
Default entry point for any research request — a hybrid router that classifies the question deterministically and either delegates to a specialist research skill (pulse for trends/sentiment, grants for NIH funding, litreview for academic literature, syllabus for course reading, patent for prior-art + IP landscape, dossier for entity research) or runs its own plan-decompose-multi-source-search-synthesize-cite fallback workflow when no specialist matches. Always surfaces the routing decision so users can override. Triggers — "research [topic]", "look into [topic]", "what do we know about [topic]", "investigate [topic]", "find me information on [topic]", "do some research on [topic]", "I need to understand [topic]", or any research request that doesn't obviously match a more-specific specialist skill. Output is a markdown briefing (default) or .docx document (on request) with full citations and an audit log.
pulse
Multi-source recency research skill that takes the pulse of any topic across Reddit, Hacker News, the open web, and optionally X/Twitter within a configurable recent window (default 30 days). Forcing intake clarifies topic specificity, angle (trend/sentiment/problems/opportunities/comparison), time window, and platform scope before searching. Returns a synthesized briefing with citations, engagement metrics, and cross-platform pattern analysis. Triggers: 'pulse on [topic]', 'what's happening with [topic]', 'what are people saying about [topic]', 'current conversation about [topic]', 'take the pulse of [topic]', 'trending: [topic]', 'find me info on [topic]', or any variation requesting multi-source recency intelligence on a topic. Also use for competitor research, trend discovery, tool comparisons, and audience sentiment analysis.
patent
Patent prior-art and landscape intelligence skill — not generic patent help. Commits to one of five sub-use-cases via forcing intake (novelty search / freedom-to-operate / competitive landscape / acquisition diligence / litigation prior-art) before any search runs. Searches Google Patents, Espacenet, USPTO, and optionally Lens.org for citation-graph signals. Output is an editable Word document (.docx) with verdict, ranked closest art (claim-text extracted), CPC-class-aware landscape, family-resolved hits, geographic coverage, FTO flags where applicable, strategy recommendations, and full audit log. Triggers: 'prior art search for [invention]', 'patent search on [topic]', 'freedom to operate analysis', 'FTO for [product]', 'patent landscape for [field]', 'is [invention] novel', 'patents on [topic]', 'competitive patent analysis', 'prior art for litigation', 'patent diligence on [company]'. Produces search signal, not legal advice — always recommends consulting a patent attorney before filing or licensing decisions. Trademark, copyright, and trade-secret questions are out of scope.
notebooklm
Browser automation skill for controlling Google's NotebookLM. Handles reading and querying notebooks, adding sources (URLs, text, files, YouTube links, synthesized content), generating Studio outputs (Audio Overview, infographics, slide decks, study guides, briefing docs, mind maps, timelines, FAQs), and creating new notebooks. Triggers on any phrase involving NotebookLM — 'open NotebookLM', 'check my [name] notebook', 'pull info from NotebookLM', 'ask my notebook about X', 'add [source] to NotebookLM', 'create an infographic in NotebookLM', 'use NotebookLM Studio', 'generate a slide deck from my notebook', or any variation where the goal involves NotebookLM. Requires browser automation environment — fails gracefully when unavailable.
litreview
Academic literature orientation skill that searches papers via Consensus, builds a strategic search plan using PICO (default) or SPIDER / Decomposition / hybrid as fallbacks, and synthesizes findings into a professionally formatted Word document (.docx) research guide. Grill-me intake (research question specificity + framework hint + tentative depth) before the recon search; a second forcing checkpoint after Phase 2 confirms framework + sub-areas + depth before searches consume budget. Configurable depth (5/10/20 queries) controls coverage vs. speed. Output is a 'launching pad' — not a finished review, but an orientation guide that lets a researcher dive in confidently. Triggers: 'litreview on [topic]', 'literature review on [topic]', 'I'm starting a literature review on X', 'I'm writing a paper on X', 'help me research X', 'I'm doing research on X', 'can you help me research X'. Do NOT trigger for single one-off paper searches where the user just wants a quick list — that's a plain Consensus search.