Use cases
Execution Verification Becomes Necessary Anytime Software Execution Changes Consequential State
Execution Verification Infrastructure applies wherever supported software execution produces state changes whose execution history and state lineage matter.
Consequential execution requires a verifiable history to establish how an outcome was produced.
S1S2S3S4 Human execution Agent execution Tool execution Automation execution
Different actors. Different environments. Same systems property.
The architecture stays the same
The Environment Changes. The Requirement Remains the Same.
Different execution environments. The same fundamental need: establish what executed and how it produced the resulting state change.
Different actors. Same requirement.
Human
Agent
Tool
Automation
Runtime
Controller
The Execution Verification architectural requirement applies to consequential execution, not to a particular application.
Current Use Cases
Execution Verification Infrastructure strengthens and complements existing platform systems.
Production incident investigation
Production State Changed. What Recorded Execution Was Associated with It?
During incident investigation, teams need to work backward from a consequential state change toward the execution associated with it, to determine what caused the outcome.
S1S2S3S4 Human Agent Tool Runtime$ salmon lineage state:S4
- predecessor
- S3
- event
- 04F7A2...
- actor
- agent-a
- action
- write
- verification
- ✓ VERIFIED
AI agent generated software changes
When Agents Modify Software, Preserve the Execution Record
Agent generated changes can involve delegated execution, tools, subprocesses and multiple state transitions. Execution history provides a durable record of supported execution relationships to verify outcomes.
S1S2 E001 humanE002 agent-aE003 toolSensitive code and configuration changes
Consequential Changes Need More than a Final Diff
A diff describes how an artifact changed. An execution record preserves evidence associated with the supported execution and state transition that produced the resulting state.
Configuration change
S1
EXEC 7A21 agent-bmodify() Signed
Execution event
- artifact
- production.yaml
- actor
- agent-b
- state_before
- 91d7a4e2...
- state_after
- a4e8c1f9...
- signature
- ed25519:3a...
- verification
- ✓ VERIFIED
Delegated agent execution
AI Agent Delegation Increases the Importance of Execution History and State Lineage
As execution occurs between actors and tools, preserving actor-execution relationships and recorded state lineage becomes increasingly important.
State lineage
S1 Agent A Agent B S2 S3The architectural pattern
The Same Requirement Appears Wherever Autonomous Systems Change Consequential Software State
The following examples illustrate where Execution Verification Infrastructure becomes valuable across various environments.
Same mechanism.
Multiple environments.
Same architecture.
Autonomous security remediation
Autonomous Security Systems Are Also Software Actors
When a security agent modifies configuration, blocks execution, terminates processes or performs remediation, the controller itself is a consequential software actor.
Threat / condition Security agent S0 S1 remediate()Infrastructure automation
Machine-Driven Infrastructure Changes Can Produce High-Consequence State
Infrastructure automation increasingly allows software actors to change environments through APIs and control planes. Where execution is supported by Salmon capture, the same execution-record model applies.
Automation Cloud API Configuration RuntimeS0 S1infrastructure
executionCross-system agent + tool execution
Execution Histories Increasingly Cross System Boundaries
As agents and tools operate across systems, preserving execution history becomes more important — and execution outside the supported capture boundary remains explicit.
Supported capture
Agent Tool APIUnaccounted-for state transition External service
Different Use Cases. One Execution Verification Layer.
Software
changes
Security
remediation
Infrastructure
automation
Agent
workflows
Execution Verification Infrastructure Capture · Execution Record · State Lineage · Verification
Verifiable execution evidence
- Security
systems 
AI / Agent
systems- Engineering systems
- Audit / Governance
What Changes — and What Doesn't
Changes Across Use Cases
- Actor type
- Tool / API
- Runtime environment
- Artifact type
- Authority context
- Execution environment
- Operational domain
Remains Constant
Execution event model- Recorded state relationships
- Event linkage
- State lineage
- Verification
- Explicit boundaries / gaps
- Core architectural primitive
Execution Verification Follows
Consequential Execution.
Wherever supported autonomous software execution changes consequential state — preserving a verifiable execution record becomes a foundational systems requirement.
When AI Agents Change Software State, the History Should Be Verifiable.
A cryptographic execution record for consequential state changes.