From daily-chief
Run the user-invoked Daily Chief Plan today workflow with device login, a local preview, and explicit confirmation before remote apply.
How this skill is triggered — by the user, by Claude, or both
Slash command
/daily-chief:daily-chiefThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Remote truth. Visual preview. Explicit apply.
Remote truth. Visual preview. Explicit apply.
Use this workflow only when the user invokes /daily-chief or /skill:daily-chief. Only Plan today is implemented. Daily Chief is the remote source of truth. The local HTML file is a visual preview, never task or plan state.
Check authentication. If credentials are missing or rejected, run browser device login and wait for approval.
npx [email protected] status --json
npx [email protected] auth
Read the request and prior conversation. Ask only for missing canonical answers: available hours, energy (high, normal, or low), must happen today, fixed events, and what can defer.
Create the remote, non-mutating preview. date is optional and the server chooses its current UTC date when omitted.
npx [email protected] plan-today --answers-json '{"availableHours":4,"energy":"normal","mustHappen":"Ship the review","fixedEvents":"15:00 client call","canDefer":"Inbox cleanup"}' --json
Inspect the returned local HTML out file and its recommendations. It is local only and has no token. Do not publish it or use htmldrop.
Require clear confirmation to apply that exact preview. Viewing it, silence, or a vague acknowledgment is not confirmation.
Apply only the returned applyPayload, unchanged. It contains the stored previewId and its generated idempotencyKey.
npx [email protected] apply-plan --json '<applyPayload from plan-today>' --output-json
Report an applied plan only after the command receives a successful remote response. Never expose, log, or embed bearer tokens. For API semantics and self-host configuration, read the API reference.
npx claudepluginhub ooiyeefei/ccc --plugin daily-chiefCreates structured, bite-sized implementation plans from specs or requirements before writing code. Useful for breaking down multi-step tasks into testable steps with file structure and task boundaries.