Module dry_run 

zig run --dry-run — preview what a workflow would do without invoking the agent.

This module walks the tiers produced by topological_sort and renders, for each step, everything a real run would compute up to the moment of agent invocation: the resolved prompt, system prompt (including the <resources> / <memory> / <storage> blocks), the condition outcome, and the full AgentConfig snapshot (every AgentBuilder knob the executor would set plus any command-specific params) that would be applied.

No side effects: no session log is written, no storage directories are created, no agent subprocess is launched.

Structs

DryRunBlock

DryRunBlocks

DryRunCondition

DryRunContext

Inputs print_plan needs to build the plan. Mirrors the state execute() has assembled just before it would open a session log.

DryRunDisabled

DryRunPlan

DryRunSave

DryRunStep

DryRunTier

DryRunWorkflow

Enums

DryRunFormat

Output format for a dry-run plan.

Functions

print_plan

Build the plan and print it to stdout in the requested format.