Subagent behaviour suite for the desktop Chat: /eplus-agent-verification:run spawns the plugin's own one-line Haiku probes (plain, background, skill-forked, preloaded skill, Sonnet, maxTurns, nested, memory) and writes a verdict report; a host-side probe hook logs every Agent call and subagent start and stop with timestamps into the session export so foreground versus background is decided from evidence. Testing profiles only.
4 skills · 2 commands · 9 agents · hooks
Live acceptance suite: /eplus-acceptance:run has the model exercise every installed EPLUS plugin in this Cowork session (identity note, failure nudge, main-thread and subagent reports, routing workers, RFI read and researcher echo, presence checks) and write a verdict report with evidence into the session export. No hooks; costs nothing while idle.
1 skill · 4 commands
On-demand test suite for every hook in every installed EPLUS plugin. /eplus-hook-verification:verify-hooks replays each wiring against fixtures and expectations on the Windows host and writes the results into the session folder the export zips. Testing profiles only: its trigger hooks cost one PowerShell start per prompt and per Glob call.
1 skill · 1 command · hooks