Skip to content

armgit / ruflo

Public Imported from GitHub
0

🌊 The original agent harness. Deploy intelligent multi-player swarms, coordinate autonomous workflows, and build conversational AI systems. Features adaptive memory, self-learning intelligence, federation, vector RAG integration, and native Claude Code / Codex / Hermes and many more Integrated

checkpoint: File edit:

rUv2fd2edf61809
diff --git a/v3/mcp/tools/agent-tools.ts b/v3/mcp/tools/agent-tools.ts
index 858484d9b9..02274c0cc7 100644
--- a/v3/mcp/tools/agent-tools.ts
+++ b/v3/mcp/tools/agent-tools.ts
@@ -21,12 +21,52 @@ function generateSecureAgentId(): string {
return `agent-${timestamp}-${random}`;
}
+// ============================================================================
+// Agent Type Validation (Security: Restrict to known agent types)
+// ============================================================================
+
+/**
+ * Known agent types - validates against allowed agent types to prevent
+ * arbitrary code execution or unexpected behavior from unknown types.
+ */
+const ALLOWED_AGENT_TYPES = [
+ // Core Development
+ 'coder', 'reviewer', 'tester', 'planner', 'researcher',
+ // Swarm Coordination
+ 'hierarchical-coordinator', 'mesh-coordinator', 'adaptive-coordinator',
+ 'collective-intelligence-coordinator', 'swarm-memory-manager',
+ // Consensus & Distributed
+ 'byzantine-coordinator', 'raft-manager', 'gossip-coordinator',
+ 'consensus-builder', 'crdt-synchronizer', 'quorum-manager', 'security-manager',
+ // Performance & Optimization
+ 'perf-analyzer', 'performance-benchmarker', 'task-orchestrator',
+ 'memory-coordinator', 'smart-agent',
+ // SPARC Methodology
+ 'sparc-coord', 'sparc-coder', 'specification', 'pseudocode',
+ 'architecture', 'refinement',
+ // Specialized Development
+ 'backend-dev', 'frontend-dev', 'mobile-dev', 'ml-developer',
+ 'cicd-engineer', 'api-docs', 'system-architect', 'code-analyzer',
+ // V3 Specialized
+ 'queen-coordinator', 'security-architect', 'security-auditor',
+ 'memory-specialist', 'swarm-specialist', 'integration-architect',
+ 'performance-engineer', 'core-architect', 'test-architect', 'project-coordinator',
+] as const;
+
+type AgentType = typeof ALLOWED_AGENT_TYPES[number];
+
+const agentTypeSchema = z.enum(ALLOWED_AGENT_TYPES).or(
+ z.string().regex(/^[a-zA-Z][a-zA-Z0-9_-]*$/,
+ 'Agent type must start with a letter and contain only alphanumeric characters, hyphens, and underscores'
+ ).max(64, 'Agent type must not exceed 64 characters')
+);
+
// ============================================================================
// Input Schemas
// ============================================================================
const spawnAgentSchema = z.object({
- agentType: z.string().describe('Type of agent to spawn (e.g., coder, reviewer, tester)'),
+ agentType: agentTypeSchema.describe('Type of agent to spawn'),
id: z.string().optional().describe('Optional agent ID (auto-generated if not provided)'),
config: z.record(z.unknown()).optional().describe('Agent-specific configuration'),
priority: z.enum(['low', 'normal', 'high', 'critical']).default('normal'),