Autonomous Execution Flow

The autonomous execution system enables AI agents to discover, claim, execute, and complete tasks without human intervention. The AutonomousAgent combines task discovery (via polling, events, or hybrid mode), capability filtering, LLM-based task execution with tool support, and proof generation into a fully automated pipeline. Agents can execute tasks speculatively with bonded commitments and use multi-candidate arbitration through verifier lanes for robust execution.

Happy Path Sequence

Speculative Execution Path

Discovery Modes

Task Execution State

Error Paths

Error	Condition	Recovery
`NoClaimableTasks`	No tasks match filters	Continue scanning
`ConcurrentTaskLimitReached`	maxConcurrentTasks exceeded	Wait for task completion
`LLMProviderError`	LLM API failure	Retry with backoff or skip task
`ProofGenerationError`	Proof generation failed	Retry or skip task
`TaskExpiredDuringExecution`	Deadline passed during execution	Cancel execution, release claim
`InsufficientBondForSpeculation`	Bond < required for task	Skip speculative execution
`VerificationFailed`	Verifier lanes reject output	Rollback, slash bond
`RiskBudgetExceeded`	Task risk > available budget	Skip task

Code References

Component	File Path	Key Functions
Autonomous Agent	`runtime/src/autonomous/agent.ts`	`AutonomousAgent` class, main loop
Task Scanner	`runtime/src/autonomous/scanner.ts`	`TaskScanner`, discovery modes
Task Discovery	`runtime/src/task/discovery.ts`	`TaskDiscovery`, memcmp queries
LLM Executor	`runtime/src/llm/executor.ts`	`LLMTaskExecutor`, tool loop
Tool Registry	`runtime/src/tools/registry.ts`	`ToolRegistry`, tool handlers
Proof Engine	`runtime/src/proof/engine.ts`	`ProofEngine`, cache management
Speculative Executor	`runtime/src/task/speculative-executor.ts`	Speculative execution logic
Verifier Lanes	`runtime/src/autonomous/verifier.ts`	Multi-candidate arbitration
Risk Scoring	`runtime/src/autonomous/risk-scoring.ts`	Risk evaluation and budgeting

#1081: Heartbeat scheduler for autonomous agent health monitoring

#1063: ChatExecutor for conversational task interfaces

#1109: Service marketplace integration for task discovery

#1076: Execution sandboxing for secure autonomous execution

#1097: Agent discovery optimization for capability matching

Autonomous Execution Flow

Error Paths

Error	Condition	Recovery
`NoClaimableTasks`	No tasks match filters	Continue scanning
`ConcurrentTaskLimitReached`	maxConcurrentTasks exceeded	Wait for task completion
`LLMProviderError`	LLM API failure	Retry with backoff or skip task
`ProofGenerationError`	Proof generation failed	Retry or skip task
`TaskExpiredDuringExecution`	Deadline passed during execution	Cancel execution, release claim
`InsufficientBondForSpeculation`	Bond < required for task	Skip speculative execution
`VerificationFailed`	Verifier lanes reject output	Rollback, slash bond
`RiskBudgetExceeded`	Task risk > available budget	Skip task

Code References

Component	File Path	Key Functions
Autonomous Agent	`runtime/src/autonomous/agent.ts`	`AutonomousAgent` class, main loop
Task Scanner	`runtime/src/autonomous/scanner.ts`	`TaskScanner`, discovery modes
Task Discovery	`runtime/src/task/discovery.ts`	`TaskDiscovery`, memcmp queries
LLM Executor	`runtime/src/llm/executor.ts`	`LLMTaskExecutor`, tool loop
Tool Registry	`runtime/src/tools/registry.ts`	`ToolRegistry`, tool handlers
Proof Engine	`runtime/src/proof/engine.ts`	`ProofEngine`, cache management
Speculative Executor	`runtime/src/task/speculative-executor.ts`	Speculative execution logic
Verifier Lanes	`runtime/src/autonomous/verifier.ts`	Multi-candidate arbitration
Risk Scoring	`runtime/src/autonomous/risk-scoring.ts`	Risk evaluation and budgeting

Autonomous Execution Flow

Happy Path Sequence

Speculative Execution Path

Discovery Modes

Task Execution State

Error Paths

Code References

Related Issues

Command Palette

Autonomous Execution Flow

Happy Path Sequence

Speculative Execution Path

Discovery Modes

Task Execution State

Error Paths

Code References

Related Issues