Google Ads account audit and business context setup. Run this first — it gathers business information, analyzes account health, and saves context that all other ads skills reuse. Trigger on "audit my ads", "ads audit", "set up my ads", "onboard", "account overview", "how's my account", "ads health check", "what should I fix in my ads", or when the user is new to NotFair and hasn't run an audit before. Also trigger proactively when other ads skills detect that business-context.json is missing.

Use for natural-science or engineering tasks, scientific software routing, simulation, dataset analysis, model fitting, package checks, HPC-through-shell work, validation, and evidence-backed scientific claims using DeepScientist's `artifact.science(...)` Science Evidence Graph. Includes a progressive-disclosure catalog of FermiLink skilled-scipkg package cards.

Use this skill to cut an APM release from the current worktree: assess whether the cycle since the last tag warrants a patch or minor bump (semver discipline against the merged-since-last-tag diff), sanitize the [Unreleased] CHANGELOG block into a dated version block with one concise "so what" entry per merged PR (drop internal-only churn, consolidate duplicates), bump pyproject.toml + uv.lock, run the CI-mirror lint chain, and open the release PR. Activate on "ship a release", "cut v0.x", "release prep", "bump and PR", "open release PR", "what kind of release do we need", or any phrasing that ends in opening a release PR -- even when the user does not say "skill". Stops BEFORE tagging; tagging stays a human gate that triggers the release workflow. Refuses to bump to a major (>= 1.0.0) version without explicit operator confirmation.

Recover from missing DCO sign-off on commits. Handles the single-commit amend, older-commit recovery via interactive rebase or cherry-pick, and explains the Probot DCO check that blocks PRs without sign-off. Use when a PR fails the DCO check, when a commit needs a Signed-off-by trailer added retroactively, or when sign-off was forgotten during rebase / cherry-pick / amend.

Record an architecture decision as an ADR in docs/adr/. Use when choosing between frameworks, libraries, databases, or architectural patterns; stating a decision with reasoning ("we decided X instead of Y because..."); or querying past decisions ("why did we choose X?").

Use this skill when delegating a subtask to a sub-agent, spawning a parallel worker, or handing off work across sessions. Write a self-contained task description so the receiving agent needs no prior context.

为 MaaEnd 添加、导入或补充节点识别测试截图，并按仓库约定脱敏图片、放入 tests/MaaEndTestset、维护 tests/**/test_*.json 和运行 pnpm check/pnpm test 验证。用户提到“添加测试图片/截图”“把截图加入 ADB 或 Win32 测试集”“给节点补正例或反例”“更新 hits/box”“UID 打码后提交测试图”时使用；即使用户只说把某张游戏截图放进 MaaEnd 测试，也要用本技能避免未脱敏原图进入仓库。

Orient Claude from Basic Memory before substantial repo work by reading active tasks, open decisions, recent checkpoints, and repo conventions. Use when resuming old work or when the user asks where things stand.

Capture a durable engineering decision in Basic Memory with rationale, alternatives, consequences, and affected work. Use when the user makes or asks to record a real choice — "let's go with X", "record this decision", or runs /basic-memory:bm-decide.

Save a deliberate work checkpoint to Basic Memory with the story, changed files, verification, decisions, blockers, and the next action. Use when the user asks to checkpoint, wrap up, hand off, or remember the state of the work.

Local-first URL fetch with clean markdown, structured metadata, JS-rendered SPA support, authenticated browser sessions, PDFs, and content change detection. Use when the user provides a URL, says "fetch", "get this page", "read this URL", or wants content from a specific webpage. Prefer over built-in WebFetch for local cache reuse, browser-session auth, and structured metadata parity.

Hybrid semantic discovery — fuses embeddings + keyword search + live web search via 3-way Reciprocal Rank Fusion. Use when the user has a good source and wants more like it, says "find similar", "related pages", "more like this", or wants to discover content related to a known URL or concept. Works best after a `crawl` or several `fetch` calls have warmed the local cache. Emits `cold_start` when local signals are weak.

Local-first structured extraction from any webpage — tables, definition lists, key-value pairs, JSON-LD, microdata, chart hints (SVG titles / aria-labels / figcaptions), brand assets, and metadata. Use when the user wants structured data, pricing tables, feature comparisons, or says "extract the table", "get structured data", "pull the pricing", "extract as JSON". For autonomous navigation across many pages, use wigolo's `agent` tool instead.

Compare two versions of a page and see exactly what changed — a live URL against its cached copy, two URLs, or two markdown blobs. Section-level hunks, word- or line-level granularity, or a summary of counts. Use when the user says "what changed", "diff these", "compare this page to last time", "show me the changes", "did this doc update", or wants a changelog between two versions of the same content.

Master HTTP QA: chains devenv then interactive API key / X-Environment-ID then real curls. Shards suites to parallel workers. Trigger: apitest, api test, curl localhost, QA local API.

Split a C# file into Unity 6.5+ and pre-Unity 6.5 variants. Use when a file needs different implementations for different Unity versions due to API changes (e.g., EntityId vs int, GetEntityId vs GetInstanceID).

Build local constrained-browser agents with a safe_browser tool that owns CDP, enforces a domain allowlist with Fetch interception, and lets a runtime Claude Agent SDK agent complete browsing tasks without raw browser, shell, or CDP access. Use when the user wants an agent to browse or scrape while staying on approved domains, demo blocked off-domain navigation, or generate a safe browser client.

Respond to all pending review comments on the current PR — fetch comments, apply fixes, verify accuracy, test, commit, and reply. Use when addressing Copilot reviews, GitHub PR reviews, or any batch of review feedback.

Open or create a repo-native visual documentation workspace backed by local Plan MDX files. Use when the user asks to visualize a repository, create durable visual docs for APIs/components/models/flows, launch a visual repo viewer, review repo docs like a visual IDE, or collect Plan comments that should become coding-agent changes.

Use when the user wants to interact with Lark/Feishu in any way — including but not limited to: reading, querying, searching, sending, replying to, forwarding, or downloading Lark/Feishu IM messages and chat history; managing group chats; listing, viewing, searching, creating, or editing cloud docs/files, Drive, Markdown, spreadsheets, Base tables, wiki, whiteboards, slides, apps, calendars/events/meeting rooms, contacts/org structure, tasks/todos, approvals, attendance, mail, minutes, VC notes, OKRs, real-time events, or custom Lark CLI skills. Load this skill first to find the right sub-skill.