feat(mcp): add Redis MCP Bridge with 25 tools (#2511)

[mrveiss]

Summary

Implements redis_mcp as the 11th native MCP bridge in AutoBot's registry, giving AI agents direct access to Redis data structures, RediSearch vector search, and operational intelligence.

Closes #2511

25 MCP Tools across 3 categories:

Data Access (15 tools):

• String: redis_get, redis_set, redis_delete
• Hash: redis_hget, redis_hgetall, redis_hset
• List: redis_lrange, redis_lpush, redis_rpush
• Sorted Set: redis_zrange
• Stream: redis_xrange, redis_xadd
• Keys: redis_scan_keys, redis_type, redis_ttl

Vector Search (4 tools):

• redis_vector_create_index, redis_vector_search, redis_hybrid_search, redis_vector_index_info

Ops Intelligence (6 tools):

• redis_server_info, redis_dbsize, redis_memory_stats, redis_stream_health, redis_client_list, redis_slowlog

RBAC Model:

• Users: Read all + write autobot:agent:* namespace only, read-only ops (no client_list/slowlog)
• Admins: Full access, destructive ops (delete) require approval

Security hardening (code review findings):

• All keys validated in namespace check (not just the first) — prevents multi-key bypass
• filter_expression sanitized via regex allowlist — prevents RediSearch injection
• top_k clamped to [1, 500] — prevents resource exhaustion
• create_index restricted to SCOPED_WRITE for non-admin users
• database parameter validated against DATABASE_MAPPING

Architecture:

• All tools use autobot_shared.redis_client (async, pooled, circuit-broken)
• Registered in both MCP registry (mcp_registry.py) and router registry (core_routers.py)
• Follows existing MCP bridge patterns (filesystem_mcp, database_mcp)
• Agent system prompts auto-discover Redis tools via _get_mcp_tools_prompt() (Bug: MCP dispatch fallback doesn't forward user role for RBAC filtering #2629)

Files:

• autobot-backend/api/redis_mcp/__init__.py — Package entry
• autobot-backend/api/redis_mcp/tools.py — 25 tool definitions (JSON schemas)
• autobot-backend/api/redis_mcp/rbac.py — Role-based access control matrix
• autobot-backend/api/redis_mcp/data_access.py — 15 data tool handlers
• autobot-backend/api/redis_mcp/vector_search.py — 4 vector search handlers
• autobot-backend/api/redis_mcp/ops_intelligence.py — 6 ops tool handlers
• autobot-backend/api/redis_mcp/router.py — FastAPI router + dispatch

Test plan

• Verify all 25 tools listed at GET /api/redis/mcp/tools
• Verify RBAC: user cannot access redis_client_list or redis_slowlog
• Verify namespace enforcement: user write to non-autobot:agent:* key is rejected
• Verify multi-key namespace enforcement: all keys in a delete call are validated
• Verify admin destructive ops return approval_required status
• Verify data access tools (get/set/hash/list/stream) work against Redis
• Verify vector search creates index and queries return results
• Verify top_k is clamped (passing 9999 returns max 500)
• Verify invalid filter_expression is rejected
• Verify invalid database parameter returns 400
• Verify ops tools return valid server stats
• Verify bridge appears in MCP registry at GET /api/mcp/bridges
• Pre-commit hooks pass (black, isort, flake8, bandit, function length)

[github-actions]

⚠️ SSOT Configuration Compliance: Violations Found

Metric	Count
Total Violations	2
SSOT Violations (high priority)	1
Other Violations	1

⚠️ 1 values have SSOT config equivalents!

These should be replaced with SSOT config imports:

Python:

from src.config.ssot_config import config
# Use: config.vm.main, config.port.backend, config.backend_url

TypeScript:

import config from '@/config/ssot-config'
// Use: config.vm.main, config.port.backend, config.backendUrl

📖 See SSOT_CONFIG_GUIDE.md for documentation.