/
/
/
1# MCP - Server Configuration & Integration
2
3## SERVER CATEGORIES
4**Core Development**: git, serena, ide | **Documentation**: context7, readwise, sequential-thinking | **Database**: supabase | **Testing**: playwright, puppeteer | **Monitoring**: sentry, gmail | **Deployment**: vercel
5
6## AGENT-MCP COORDINATION
7
8### Utility Agent Delegations (MANDATORY)
9- **file-creator**: File operations (Write, MultiEdit, Edit)
10- **git-workflow**: Version control (git.git_commit, git.git_add, git.git_status)
11- **knowledge-fetcher**: Research (readwise, context7, WebSearch)
12- **date-checker**: Temporal calculations (date commands, filtering)
13- **context-fetcher**: Documentation (Read, Glob, WebFetch)
14
15### Specialized Agent Patterns
16- **Testing**: test-writer-fixer â playwright/ide â validation
17- **Code Analysis**: Engineering agents â serena/sequential-thinking â insights
18- **Error Resolution**: backend-architect â sentry/supabase â diagnosis
19- **Deployment**: devops-automator â vercel/git â monitoring
20
21## QUERY CLASSIFICATION RULES
22
23### Task Type Detection
24```yaml
25simple_lookup:
26 indicators: ["find", "get", "show", "list", "what is", "when did"]
27 rule: "Use most direct tool, STOP after definitive answer"
28 max_tools: 1-2
29
30complex_analysis:
31 indicators: ["analyze", "compare", "synthesize", "recommend", "explain why", "how should"]
32 rule: "Multi-tool coordination acceptable, sequential-thinking encouraged"
33 max_tools: 3-5
34```
35
36## ERROR RECOVERY & FALLBACKS
37- **git** â Manual git commands â Note version control limitations
38- **context7** â WebSearch â Manual documentation lookup
39- **sequential-thinking** â Native analysis â Note complexity limitations
40- **puppeteer/playwright** â Manual testing â Provide test cases and fallback instructions
41- **serena** â Text-based code analysis â Note semantic analysis limitations
42- **sentry** â Manual error tracking â Log analysis and issue documentation
43- **supabase** â Manual SQL operations â Database connection alternatives
44- **vercel** â Manual deployment â CI/CD pipeline alternatives
45- **readwise** â Manual search â Note knowledge management gaps
46- **gmail** â Manual email operations â Note communication workflow disruption
47- **ide** â Text-based diagnostics â Note development environment limitations
48
49## ANTI-PATTERNS & USAGE WARNINGS
50
51### Critical Avoid Conditions
52- **supabase_operations**: read_only_environment, production_lockdown
53- **browser_automation**: headless_server_environment, rate_limited_apis
54- **sentry_operations**: insufficient_permissions, service_outage
55- **git_operations**: detached_head_state
56- **sequential_thinking**: simple_single_step_tasks
57
58### Performance Anti-Patterns
59- **excessive_sequential_thinking_calls**: >3 calls per task
60- **browser_automation_for_api_tasks**: Use direct API calls when available
61- **readwise_export_overuse**: STOP after list_highlights unless full content specifically needed
62- **unnecessary_tool_escalation**: Apply STOP rule from query classification
63
64## PERFORMANCE OPTIMIZATION
65
66### Stop Conditions
67```yaml
68stop_execution_when:
69 query_answered: "Got definitive answer to user's specific question"
70 token_threshold_reached: "simple_lookup: >1000 tokens, medium_analysis: >5000 tokens, complex_analysis: >15000 tokens"
71 completion_criteria_met: "Task objectives fully satisfied"
72 error_resolution_complete: "Problem identified and solution provided"
73```
74
75### Tool Performance Profiles
76- **git_operations**: 50-200ms, minimal tokens, parallel_safe
77- **supabase_operations**: 200-2000ms, low-medium tokens, NOT parallel_safe
78- **sentry_operations**: 300-30000ms, low-very_high tokens, parallel_safe
79- **browser_automation**: 800-5000ms, medium tokens, resource_intensive, NOT parallel_safe
80- **sequential_thinking**: 1000-10000ms, high-very_high tokens, parallel_safe
81- **serena_operations**: 100-2000ms, low-medium tokens, parallel_safe
82- **readwise_operations**: 300-10000ms, low-very_high tokens, parallel_safe