feat: add Memory MCP server allow list to repository settings (#203)

Add all 9 Anthropic Memory MCP server tool calls to the repository-level
allow list (.claude/settings.json) to enable Claude Code agents to use
memory server features without requiring user approval.

Memory MCP calls added:
- mcp__memory__create_entities
- mcp__memory__create_relations
- mcp__memory__add_observations
- mcp__memory__delete_entities
- mcp__memory__delete_observations
- mcp__memory__delete_relations
- mcp__memory__read_graph
- mcp__memory__search_nodes
- mcp__memory__open_nodes

This enables full knowledge graph functionality (create, read, update,
delete) for storing architectural decisions, patterns, and learnings
across development sessions.

Aligns with project principle: "Proactively use Memory MCP server to
store architectural decisions and learnings"

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-authored-by: Claude Sonnet 4.5 <noreply@anthropic.com>

Author: jeremyeder

.claude/settings.json

@@ -0,0 +1,25 @@
+{
+  "permissions": {
+    "allow": [
+      "Skill(frontend-design:frontend-design)",
+      "Bash(black:*)",
+      "Bash(isort:*)",
+      "Bash(jq:*)",
+      "Bash(env)",
+      "Bash(harbor:*)",
+      "Bash(agentready assess:*)",
+      "Bash(pytest:*)",
+      "mcp__memory__create_entities",
+      "mcp__memory__create_relations",
+      "mcp__memory__add_observations",
+      "mcp__memory__delete_entities",
+      "mcp__memory__delete_observations",
+      "mcp__memory__delete_relations",
+      "mcp__memory__read_graph",
+      "mcp__memory__search_nodes",
+      "mcp__memory__open_nodes"
+    ],
+    "deny": [],
+    "ask": []
+  }
+}