New in week 43

Convene the Council of High Intelligence — multi-persona deliberation with historical thinkers for deeper analysis of complex problems.

Analyzes Move language packages against the official Move Book Code Quality Checklist.

UI animation rules from Emil Kowalski's "Train Your Judgement" (emilkowal.ski/ui/train-your-judgement).

Swift Design Systemを使用したiOS UI実装スキル。デザイントークン（カラー、タイポグラフィ、スペーシング、角丸、モーション、エレベーション）、UIコンポーネント（Button、Card、Chip、FAB、Snackbar、Picker）、レイアウトパターン（AspectGrid、SectionC…

Understand an existing feature end-to-end and produce a functional spec in /specs plus a user-facing doc in /docs.

Drive Claude Code task-by-task with test gates.

Generate TASKS.md + milestones with verification gates.

Build the proof-pack folder (manifest + reports + inventory).

Run deterministic test gates and summarize results.

Resource-first knowledge vault for an AI agent.

Use when a coding request contains work you can write a tight spec for and verify from a diff: boilerplate, scaffolding, repetitive edits, test stubs and…

Use when executing a multi-task plan or a build that should be driven off the paid Claude session: one Opus orchestrator routes each task by complexity…

Read-only skill for inspecting, re-anchoring, and comparing saved source materials when the current task depends on source_ref continuity.

Workspace-focused skill for inspecting files, running bounded regression checks, and continuing from the same filesystem artifact surface.

Create, track and update team goals via the Deeplake virtual filesystem at memory/goal/.

Query the local code graph (functions, classes, calls, imports) through the Deeplake mount at memory/graph/.

Global team and org memory powered by Activeloop.

Query the local code graph (functions, classes, calls, imports) through the Deeplake mount at memory/graph/.

Create, track and update team goals in Hivemind via the `hivemind` CLI.

Global team and org memory powered by Activeloop.

Use Helium Harness for browser interaction in the Helium browser.

Use when the user explicitly cannot review the output in the current domain, says they do not understand it or will copy-paste without checking, or requests…

Use when the user asks what they are not seeing, what they forgot, what could go wrong, or what an expert would say about a plan, decision, document, code…

Use when the user asks how clueless works, what its commands or levels are, or types /clueless-help.

Search, compare, evaluate, and hand off hotel bookings using the MoodTrip.ai MCP server (api.moodtrip.ai).

Use when you need a proof-of-loss cover letter or a supplement request to a carrier.

Use when someone asks what clock applies to a claim in a state.

Use when a new loss call or email comes in and you want to capture everything before it is lost.

Use when you are back from a site visit and want field notes and a photo list turned into a dated inspection note.

Use when a carrier estimate comes back and you want to know what to challenge.

Use when someone pastes meeting notes, a transcript, or a thread and asks what was decided.

Use when a meeting just ended and someone asks for the follow-up, recap, or next-steps email.

Use when someone shares an upcoming meeting, an agenda, or a calendar invite and wants to walk in prepared.

Use when someone dumps their task list, inbox, or worries and asks what to work on this week.

Use when someone must tell customers about a release, feature, policy change, or price change and asks for the announcement, release note, or email.

Use when a customer or client wrote in and someone asks for a reply, a response, or what to say back.

Use when a draft runs long and someone asks to cut, shorten, tighten, or trim it, or get it down to half.

Use when a draft sounds like an AI wrote it, or someone asks to make it less AI, more human, plain English, or like them.

Use when a payment failed or an invoice is past due and you want a personal note, not the processor's robot.

Use when a customer just canceled and you want to learn why without a survey.

Use when a customer is about to cancel, went quiet, or asked to downgrade, and you need to decide what to say.

Use when a pack is ready to list and someone asks for the README, store page, or buyer summary.

Use when a new pack version ships and someone pastes a diff or a list of changes and asks for release notes.

Use when a skill won't fire, or someone asks if a description matches what buyers type.

Use when someone wants to turn a process into a skill, or asks how to write a SKILL.md.

Address GitHub PR feedback by adversarially verifying comments, fixing valid issues, replying, and resolving threads.

Review a GitHub PR and prepare prioritized comments for submission, with optional approval or request for changes.

Adversarially audit staged, unstaged, and untracked changes before a commit or PR, and fix clear low-risk issues.

Create a handoff prompt for a future AI session when resetting a thread or transferring work.

Organize Firebase Crashlytics crash issues into evidence-backed per-issue folders and generate incremental crash reports that compare with the previous report.