Spark CLI Skills 1.2.2

Updates the use-spark base skill to Spark CLI 1.2.2, brings the Spark CLI to Windows, and ships a new recipe.

Platform

• The Spark CLI is now supported on Windows, in addition to macOS.

Changed commands

• draft - added --attach-stream to attach a file piped from stdin (the inbound twin of attachment --stream), for files the app can't read from disk. The output now surfaces a Spark deep link that opens the draft directly.
• comment - added --attach-stream to attach a stdin-piped file as its own comment message.

New skills

• recipe-unsubscribe-audit - full unsubscribe audit: scan inbox, archive, and GateKeeper for newsletter senders, classify by volume and engagement, flag phishing and duplicate subscriptions, then unsubscribe via spark action or extracted footer links. Contributed by @quser.

Spark CLI Skills 1.2.1

Updates the use-spark base skill to Spark CLI 1.2.1

New commands

• attachment - read a single email attachment by its ID (pk) from the thread; auto-downloads if not cached. --stream for sandboxed agents that can read stdout but not local paths.
• templates - list saved message templates (personal and team) for use with draft --template.
• template - show a single template by ID or name: scope, recipients, subject, body, attachments, and its placeholder requirements.

Changed commands

• search - now has two modes. With a topic: unchanged hybrid keyword + semantic search (full bodies, top 20). Without a topic (new list mode): paged compact table of every email matching --filter / --in across all folders and accounts, newest first.
• draft - added team-shared drafts and collaboration flags: --team, --user, --remove-user, --allow-send / --no-allow-send, and --unshare. Added template support: --template and --placeholder. --body is now optional when a template supplies one.
• thread - now accepts a Spark deep link (dpl/bl, readdle-spark://, readdlespark://) in addition to a numeric message ID.
• meeting - now accepts a Spark deep link in addition to a numeric meeting ID.

Spark CLI Skills 1.1.0

Adds triage-level skills for Spark CLI 1.1.0, covering drafts, team comments, and email/contact actions, plus a new persona library.

What's new

Updated base skill

• use-spark - expanded to cover Spark CLI 1.1.0 triage-level write commands (drafts, team comments, email and contact actions) alongside the existing read flows.

New recipes (triage-level)

• recipe-inbox-zero - process the inbox to zero with category-first triage.
• recipe-vacation-catchup - process a large backlog after time away with batch processing.
• recipe-new-sender-review - process GateKeeper's new sender queue (accept or block).
• recipe-draft-batch - review and finalize multiple pending drafts in a batch session.
• recipe-meeting-followup - review a meeting transcript and draft follow-up emails with action items.
• recipe-shared-inbox-triage - process open items in shared inboxes (assign, respond, mark done).
• recipe-delegate-and-track - assign emails to teammates with context and track delegation status.
• recipe-priority-tuning - fine-tune Spark's priority category by adjusting contact-level settings.
• recipe-newsletter-cleanup - audit newsletter subscriptions (block, unsubscribe, or summarize).
• recipe-notification-hygiene - reclassify, group, and bulk-archive noisy notifications.
• recipe-label-organize - organize emails with labels and folders (attach, detach, or move).

New personas

Long-running roles that shape how the agent works across a session - tone, defaults, and which recipes to reach for.

• persona-exec-assistant - briefings, drafts, scheduling, contact lookup.
• persona-founder - high-volume triage, aggressive delegation, oversight.
• persona-team-lead - workload review, assignment distribution, standup prep.
• persona-project-manager - thread tracking, stakeholder updates, action items.
• persona-sales-rep - pipeline review, follow-up cadence.
• persona-support-agent - shared inbox processing, routing, escalation.
• persona-meeting-manager - prep, transcript review, follow-ups, scheduling.
• persona-freelancer - multi-client management, invoice follow-ups, availability.

Access levels

Per-account access is now configured in Settings → AI Agents → Spark CLI Access:

• read-only - works with the original read recipes.
• triage - everything in read-only plus drafts, comments, and email/contact actions; required by the new recipes and personas above.

Recipes and personas declare the level they need; running one against an account with insufficient access returns an error explaining how to upgrade.

Requirements

• macOS with a recent build of Spark Desktop, signed in to at least one account.
• Spark CLI enabled via Settings → AI Agents → Spark CLI Setup.

Installation

npx skills add https://github.com/readdle/spark-cli-skills

Spark CLI Skills 1.0.0

Initial release of Spark CLI Skills — AI agent skills for the Spark CLI.

What's included

Base Skill

• use-spark — Complete command reference for the Spark CLI (mail, calendar, contacts, teams, meetings).

Recipes

• recipe-inbox-by-category — Process inbox in priority order by smart category.
• recipe-morning-standup — Today's events + unread by category + team assignments.
• recipe-meeting-prep — Gather context for an upcoming meeting.
• recipe-invitation-manager — Process pending invitations with availability checks.
• recipe-schedule-meeting — Find mutual availability and suggest times.
• recipe-end-of-day — Close out the day, review loose ends, preview tomorrow.
• recipe-weekly-digest — Weekly summary of events, email, and team status.
• recipe-multi-account-review — Review inbox status across multiple accounts.
• recipe-topic-timeline — Build a chronological narrative from meetings and emails.
• recipe-stakeholder-brief — Compile a relationship dossier from all interactions with a person.
• recipe-shared-inbox-status — Review shared inbox health (open, done, unassigned items).
• recipe-team-workload — Audit team assignment distribution and spot imbalances.
• recipe-calendar-audit — Analyze meeting load, back-to-back chains, and free blocks.

Requirements

• macOS with a recent build of Spark Desktop, signed in to at least one account.
• Spark CLI enabled via Settings → AI Agents → Spark CLI Setup.

Installation

npx skills add https://github.com/readdle/spark-cli-skills