Pivot
>-
Delegate
Run the Planner→Generator→Evaluator loop — Claude (Fable) specs the work, Codex GPT-5.5 implements it, an Opus 4.8 subagent grades the result against the spec. Use for implementation tasks big enough to spec — new features, refactors, migrations, bulk mechanical changes. Not for small edits (<~30 lines) or pure judgment tasks; do those directly.
Press Release Faq
>
Vibe Security Audit
>
Validate
>-
Spring Cleaning
>
Search X
Find a specific remembered X post from fuzzy clues and verify it on the live post. Use when exact wording or the author is unknown.
Post Walk
Process a local Apple Voice Memo into an archived transcript, project notes, tasks, todos, and drafts. Use only when explicitly asked to process a walk or commute memo.
Explain
Walk the user through code changes made by coding agents so they understand what changed, why decisions were made, what trade-offs were accepted, what alternatives were or could have been considered, and what risks or follow-up choices remain. Use when the user invokes /explain, asks to understand an agent-made diff, branch, PR, commit, worktree, implementation, refactor, migration, feature, bug…
Save2md
Save articles, X posts or threads, and YouTube transcripts as local Markdown with bundled tools. Use when the user asks to save or archive a URL as Markdown, including paywalled articles.
Continue
>-
Plan
>-
Start
>-
Audit And Improve
>-
Fast Wifi
Diagnose a slow macOS network with the bundled measurement script. Use when the user asks to troubleshoot current Wi-Fi or internet performance.
Capture Design
Save reference visuals as a reusable design-system swipefile. Use when the user explicitly asks to capture or preserve an aesthetic for reuse.
Status
>-
Name It
Run a structured brand or product naming and validation process. Use for explicit naming strategy or candidate evaluation, not ordinary code identifiers.