
Scope Guard
This is for when you're orchestrating multiple AI agents and need to prevent them from stepping on each other's toes. It gives you six tools to define what each agent can and can't do: create_scope_contract sets boundaries like allowed tools and file access patterns, validate_action checks if a proposed action is within scope before execution, and detect_delegation_loop catches circular delegation chains. You also get logging, compliance reports showing violation counts and scope utilization percentages, and work duplication detection across agents. Reach for this when you're building agentic systems where agents might delegate to each other and you need guardrails to prevent chaos.
- tools
- 6
- transport
- STDIO
- last updated
- Apr 24, 2026
Tools
6 toolsPublic tool metadata for what this MCP can expose to an agent.
create_scope_contractDefine agent boundaries: allowed/forbidden tools, file access patterns, delegation limits
Define agent boundaries: allowed/forbidden tools, file access patterns, delegation limits
No parameter schema in public metadata yet.
validate_actionPre-flight check — is a tool call or file access within scope?
Pre-flight check — is a tool call or file access within scope?
No parameter schema in public metadata yet.
detect_delegation_loopAnalyze delegation chains for circular patterns
Analyze delegation chains for circular patterns
No parameter schema in public metadata yet.
log_agent_actionTrack agent actions for compliance monitoring
Track agent actions for compliance monitoring
No parameter schema in public metadata yet.
get_compliance_reportPer-agent summary: actions, violations, delegation count, scope utilization %
Per-agent summary: actions, violations, delegation count, scope utilization %
No parameter schema in public metadata yet.
detect_work_duplicationFind agents doing the same work across action logs
Find agents doing the same work across action logs
No parameter schema in public metadata yet.
Sponsors
Readme
scope-guard-mcp
MCP server that enforces multi-agent scope boundaries. Prevents scope drift, work duplication, and infinite delegation loops in agentic systems.
Tools
| Tool | Purpose |
|---|---|
create_scope_contract | Define agent boundaries: allowed/forbidden tools, file access patterns, delegation limits |
validate_action | Pre-flight check — is a tool call or file access within scope? |
detect_delegation_loop | Analyze delegation chains for circular patterns |
log_agent_action | Track agent actions for compliance monitoring |
get_compliance_report | Per-agent summary: actions, violations, delegation count, scope utilization % |
detect_work_duplication | Find agents doing the same work across action logs |
Resources
| URI | Description |
|---|---|
scope-guard://contracts | All active scope contracts |
Install
npm install
node src/index.js
Claude Desktop
{
"mcpServers": {
"scope-guard": {
"command": "npx",
"args": ["scope-guard-mcp"]
}
}
}
License
MIT
Details
- Repository
- github.com/mdfifty50-boop/scope-guard-mcp
- Category
- AI & LLM Tools
- Package
scope-guard-mcp- Transport
- STDIO
- Registry
- MCP registry, active