Where the work went

How it works

One loop runs every working day. Claude Code does the reading and sorting through skills and connectors, and every outcome comes back to me to approve. Each number links to its chart below.

Darshandecides and approvesClaude Codereads, sorts, drafts, filesSKILLSdaily-planning32humanizeiq-sdlc309pr-review89MCP CONNECTORSMicrosoft 365348 tasksSlack34 tasksClaude in Chrome309 ticketsMetabasequeriesAura2 tasks386tasks surfaced ↓309tickets created ↓88PRs reviewed ↓32days planned ↓every outcome comes back to Darshan for reviewDarshandecides and approvesClaude Codereads, sorts, drafts, filesSKILLSdaily-planning32humanizeiq-sdlc309pr-review89MCP CONNECTORSMicrosoft 365348 tasksSlack34 tasksClaude in Chrome309 ticketsMetabasequeriesAura2 tasks386tasks surfaced ↓309tickets created ↓88PRs reviewed ↓32days planned ↓

Tasks

Every task the daily sweep surfaced, and how many became a ticket on HumanizeIQ. A task counts as ticketed only when a snapshot records its ticket ID.

Skills

Who wrote each skill, and what it produced between the window dates. Only output saved to files is counted. Bars use a log scale.

SkillOriginOutputCount

MCPs

ConnectorWhat it didStatus

PR reviews

Categories

Sub-categories

By day

Tasks surfaced on each planning day, stacked by category. Hover a column for its total.

Projects

Tools found in the products and brainstorming folders. A tool is listed only when a file proves it: a manifest, AI Studio metadata, a commit line, or a plan.

Built withRuns onEvaluated
ProjectToolsEvidence

How this was checked