wan-huiyan

Orchestrate a multi-agent adversarial review panel where several Claude Code subagents with different perspectives independently review a piece of work, debate with each other, reach (or fail to reach) consensus, then a supreme judge renders the final verdict. Use this skill whenever the user asks for a "review panel", "multi-agent review", "adversarial review", "have agents debate this", "review

Integrate structured review panel findings into an implementation plan document. Takes output from agent-review-panel (or any structured review with severity-rated findings) and cross-references each finding against the plan, classifies it into an action category, applies concrete edits, and produces a traceability summary. Trigger when the user says "update the plan with review findings", "incorp

Adds a cute, hand-authored SVG favicon to a web app, the tiny icon displayed in the browser tab. Use when the user requests a favicon, browser tab icon, page logo, or when the tab shows a blank/default icon.

Integrate structured review panel findings into an implementation plan document. This skill processes structured review outputs, cross-referencing findings against the plan, classifying them into action categories, applying edits, and generating a traceability summary, triggering with phrases like "update the plan with review findings".

Measure the causal impact of a marketing campaign, promo, or intervention on a business metric (revenue, conversions, transactions) using Bayesian structural time series. Use this skill whenever the user mentions "causal impact", "campaign uplift", "promo effect", "incrementality", "did the campaign work", "revenue lift from campaign", "measure uplift", "what was the true effect", "counterfactual

Fetch Google Trends data (search interest over time) via Claude-in-Chrome browser automation, including the multi-chunk daily-resolution stitching trick for date ranges longer than 90 days. Produces calibrated daily time series with provenance files. Use this skill whenever the user wants Google Trends data, search interest data, trend_category or trend_brand_share covariates, or mentions pytrends

Use when executing implementation plans with independent tasks in the current session

Create polished, activation-focused HTML presentation slides for proposing ML/data science enhancements to non-technical clients. Starts with a brainstorming discovery phase to understand audience, decision context, and stakeholder workflows before building anything. Use this skill whenever the user asks to create a presentation, slide, deck, or proposal for a client based on analysis results, mod

End-of-session handoff that captures session knowledge, dispatches output across the canonical 7-bucket docs/ taxonomy (decisions/runbooks/analysis/references/reviews/handoffs/deliverables — aligned with memory-hygiene v3.3), triggers a doc-freshness reverse-lint + skill-freshness audit to catch stale normative guidance, emits the future-to-do plan's follow-up items as GitHub issues, updates memor

Use when orchestrating long-running autonomous work — overnight runs, multi-hour research, multi-track experiments, 10+ hour jobs — where a single Claude context window cannot hold the full run. Establishes a lean parent orchestrator, file-first subagents, and successor-handoff (fresh subagent spawned mid-run when context pressure rises). This is distinct from session-handoff, which is for end-of-

Before starting a large-scale redesign (10+ PRs that rewrite shared files like templates, base layouts, or central views), audit ALL unmerged feature branches for commits that touch the same files. Use when: (1) the user asks for a multi-PR redesign / restructure / migration, (2) the worktree is at the top of main but other long-running branches exist with active work, (3) the redesign will replac

Run an overnight autonomous work session that produces a client delivery package the next morning, with a multi-agent review panel built in to catch factual errors before they reach the client. Use when: (1) you have client deliverables that need polishing/regenerating before a morning hand-off, (2) you want a thorough quality gate without blocking on synchronous review, (3) the work scope is boun

Execute an overnight autonomous B-vs-C parallel insight-discovery workflow to uncover surprising findings from client data. This is ideal for clients seeking novel insights, for balancing LLM creativity with deterministic rigor via parallel tracks, and for tasks fitting an 8-hour autonomous window.

This skill runs an overnight autonomous workflow to resolve a cluster of 6-15 related GitHub issues, typically P1 review-panel findings, delivering them as merged PRs by morning. It is ideal when issues naturally split into two PRs and the user will be unavailable.

A specialized code-reviewer subagent, often provisioned without a Bash tool, fails to see code in unattended overnight reviews, leading to incorrect 'clean' assessments and the shipment of real bugs.

This Fire-ASAP orchestrator pattern manages multi-track autonomous workflows via scheduled triggers, not in-session agents. It's ideal for overnight processes where fixed timers waste time if tracks finish early, allowing consolidation as soon as all parallel tracks complete.

Calibrate review intensity per-PR when orchestrating a long chain (10+ PRs) of independent tasks overnight from a single implementation plan. Use this when running `superpowers:subagent-driven-development` that produces many independent PRs and the user has authorized auto-merge on green.

Use when the user wants to switch Claude Code billing between API-key (Console) and subscription (Pro/Max). Toggles `apiKeyHelper` and `env.ANTHROPIC_API_KEY` in ~/.claude/settings.json on/off. The key itself lives in macOS Keychain (service=claude-code-api-key, account=anthropic), never in plaintext on disk. TRIGGER PHRASES (any of these should activate this skill): "turn API billing on", "turn A

Scout, document, and verify external data sources — both open/free and paid/licensed. Produces comprehensive reports with field-level schemas, API response structures, code snippets, and freshness rankings through a 5-phase workflow: brainstorm, research, build, verify, correct. Use this skill whenever the user wants to discover what external data is available, document data sources for a pipeline

Plan and build dbt models from schema requirements, dashboard specs, or reporting briefs. Use this skill whenever: (1) someone provides a target schema, CSV spec, or dashboard requirements and asks to build a dbt model, (2) someone needs to map required dimensions and metrics to existing source tables in a dbt project, (3) someone asks "which source tables should I use for this report/dashboard/mo

Assess whether a training window can be extended by adding a new data source (e.g., CRM-only features to bypass behavioral data limitations). Use when: (1) the user asks "can we train on more data?", "how far back can we go?", "can we extend the training window?", or "our model only has N months of data", (2) a multi-output model has different effective training windows per output due to dynamic l

Generate pixel art SVG illustrations for READMEs, docs, or slides. Use when user says "画像素图", "pixel art", "make an SVG illustration", "README hero image", or wants a cute visual.

GitHub auto-close on squash-merge only catches ONE issue per PR even when the PR title/body says "Closes #447, #448, #449, #450". The first issue closes; the rest stay OPEN. Use when: (1) you opened a PR with a title or body listing multiple "Closes #X, #Y, #Z" references, (2) the PR squash-merged successfully (`gh pr view N --json state` = MERGED), (3) `gh issue list --state open` shows some of t

Fix "untracked working tree files would be overwritten by merge" errors when pulling after a squash merge. Use when: (1) `git pull` fails listing files that should already be on main, (2) you just squash-merged a PR and can't pull on another branch, (3) files from a merged branch appear as "untracked" blocking checkout or pull. The root cause is that squash merge creates new commits — the original

Diagnose and recover follow-up commits that were pushed to an open PR's branch but never reached `main` because the user squash-merged the PR between pushes. Use when: (1) you pushed commit 1, opened a PR, then iteratively pushed commits 2, 3, ... in response to user feedback, (2) you go to merge / verify and find the PR is already MERGED, (3) the squash commit on main has insertion counts / file

Recover from the trap where deleting a base PR's branch (via `gh api -X DELETE refs/heads/<branch>`) auto-closes any open dependent stacked PR, and the closed PR cannot be reopened or retargeted. Use when: (1) you set up a stacked PR pair (PR2's `base` field = PR1's branch instead of `main`), (2) you merged PR1 via squash, (3) you then deleted PR1's remote branch via the API (because `gh pr merge

Triage a batch of candidate data signals (from client emails, meeting notes, CRM field lists, funnel definitions) into GO/DEFER/NO-GO decisions based on data availability checks. Produces a signal inventory table and hands off GO candidates to ml-feature-evaluator for deep diagnostic. Triggers when the user asks any of these (or close variations): - "The client sent an email with these fields..."

Publish a Claude Code skill to GitHub as a polished, adoptable open-source repo, AND diagnose `claude plugin install` failures on a published skill. Use when the user says "publish this skill", "put this on GitHub", "share this skill", "release this skill publicly", "open source my skill", "make this skill installable", "create a GitHub repo for my skill", "package this skill for the marketplace",

This end-of-session handoff captures session knowledge, dispatches output across a 7-bucket taxonomy, triggers document and skill freshness audits, emits follow-up items as GitHub issues, and updates memory.