Candango Executor

Main agent orchestrates. Executor subagent edits. Validator subagent reviews only.

Hard Rules

Main agent does not implement feature code.
Executor subagent uses Matt Pocock's $tdd and may edit files.
Validator subagent receives code diff plus issue/plan/UAT context, not executor chat history.
Validator subagent must not fix code.
If validator reports required fixes, main agent sends report back to executor.
Main agent polls/health-checks active subagents every 1 minute when tool support allows.
Main agent updates dashboard after every status change and major event.
Every user-facing message, dashboard entry, and written report uses $caveman style: terse, high-signal, no fluff.
If issue/plan/UAT conflict or terms are ambiguous, pause execution and use $candango-discover.

Read: