claude-operator-skills

Keep a live interactive chat session coming back to itself, visibly, after a usage/session cap interrupts it, never a silent background process. Invoke with /qq-cap-resume, or say “keep this session alive,” “resume yourself after the cap,” “don’t go dark on me,” any time a long interactive burn needs to survive a 5-hour session cap or ride out to a weekly reset without leaving you staring at a dead thread.

The unbounded end of the AFK spectrum: spend whatever it takes for quality, budget isn’t the constraint. Stand up a self-healing, fully-autonomous “kitchen-sink max-effort” campaign for a big multi-day initiative: resilience first, creative meta strategy, workflows spawning workflows, audit-gated quality, a phone-glanceable status surface with a proud wins list, and a long living tasklist. Invoke when you say “go HAM”, “go-ham”, “go afk ham”, “qq-go-afk-ham”, “kitchen sink”, “blow me away”, “use all the quota”, or hand it a large ambitious build to run autonomously for hours or days. For the frugal end, use `qq-go-afk-lean`; for the adaptive middle, use `qq-go-afk-smart`.

The adaptive middle of the AFK spectrum: picks the right model task by task, moment by moment, instead of committing to either extreme. Stand up a calm, self-healing, fully-autonomous overnight worker for a well-scoped punch list: the top-tier model writes strict specs and does quality control, cheap and free models do the grunt work, one bounded task per cycle, each one deployed and live-verified, a phone-glanceable status surface, and a durable done log. Invoke when you say “go afk smart”, “go-afk-smart”, “qq-go-afk-smart”, “go smart”, “bounded overnight”, “delegate and verify overnight”, or hand it a finite, well-specified backlog (a punch list, a rollout, a sweep) to grind down autonomously while you sleep, without the kitchen-sink fan-out of go HAM. For the unbounded end, use `qq-go-afk-ham`; for the frugal end, use `qq-go-afk-lean`.

The moment a real, avoidable mistake happens, name it plainly, root-cause the general class of it rather than just this one instance, and fix it at a level that prevents a repeat, not a note that gets written and forgotten. Then, before starting work that looks like it could repeat that shape, check the log first. A second occurrence of the same mistake means the note wasn’t enough, and it needs to become a real check instead of a reminder. Invoke with /ww-mistake-ledger, or say “log this mistake,” “close the loop on this,” “add this to the mistake log,” “check the mistake log first,” right after a preventable error, or right before starting work that resembles one already logged.

Ask what would make this fail before you build it, not after. A one-question design-time gate, imagine it’s already broken three months from now, work backward to the most likely cause, and let the answer change the plan instead of just getting noted and ignored. Works on code, an automation, a process, or a decision. Invoke with /ww-premortem, or say “what breaks this in three months,” “pre-mortem this,” “why would this fail,” “before I build this,” at the moment you’re about to commit to an approach, not after you’ve already started.

Before filling a gap in what you know with a guess, check whether the guess is load-bearing. If getting it wrong would mean redoing real work, or building on a false premise, ask a real question with concrete options instead of assuming silently. If it’s genuinely low-stakes, state the assumption out loud and keep moving rather than stalling on it. Works on any ambiguous request, with or without an AI agent involved. Invoke with /ww-questioning, or say “wait, let me ask first,” “is this load-bearing,” “check before I guess,” “clarify before building,” whenever a request has more than one reasonable reading and you’re about to just pick one.

Argue against your own plan before you build it. A five-step protocol, write the plan, critique it like an outsider would, then defend it against that critique, keep only what survives, apply the rest. Works on any plan, code or content. Invoke with /ww-steelman, or say “steelman this,” “poke holes in this,” “what am I missing,” “play devil’s advocate,” before locking in a plan, a draft, or a decision.

Serialized backend/API audit using 22 deep framework subskills. Each framework runs in its own agent context with expert-level depth. Smart interview pre-fills from conversation context.

Serialized compliance audit using 15 deep framework subskills. Each framework runs in its own agent context with expert-level depth. Smart interview pre-fills from conversation context.

Serialized content/copy audit using 22 deep framework subskills. Each framework runs in its own agent context with expert-level depth. Smart interview pre-fills from conversation context.

Serialized data quality audit using 15 deep framework subskills. Each framework runs in its own agent context with expert-level depth. Smart interview pre-fills from conversation context.

Serialized DevOps audit using 15 deep framework subskills. Each framework runs in its own agent context with expert-level depth. Smart interview pre-fills from conversation context.

Serialized frontend build quality audit using 22 deep framework subskills. Each framework runs in its own agent context with expert-level depth. Smart interview pre-fills from conversation context.

Meta-orchestrator that selects and sequences the right combination of audit domains based on project context, conversation state, and user intent. Routes to 13 domain-specific audit skills containing 255 expert-persona frameworks.

Serialized performance audit using 22 deep framework subskills. Each framework runs in its own agent context with expert-level depth. Smart interview pre-fills from conversation context.

Serialized product audit using 20 deep framework subskills. Each framework runs in its own agent context with expert-level depth. Smart interview pre-fills from conversation context.

Serialized security audit using 23 deep framework subskills. Each framework runs in its own agent context with expert-level depth. Smart interview pre-fills from conversation context.

Serialized SEO audit using 15 deep framework subskills. Each framework runs in its own agent context with expert-level depth. Smart interview pre-fills from conversation context.

Serialized QA/Testing audit using 22 deep framework subskills. Each framework runs in its own agent context with expert-level depth. Smart interview pre-fills from conversation context.

Creates and maintains a live Notion tracking doc for audit runs. Standard format with overview table, live feed, per-domain sections, and findings summary. Used by qq-audit-master automatically.

Serialized UX audit using 20 deep framework subskills. Each framework runs in its own agent context with expert-level depth. Smart interview pre-fills from conversation context.

Serialized Visual UI Design audit using 22 deep framework subskills. Each framework runs in its own agent context with expert-level depth. Smart interview pre-fills from conversation context.

Externalize-first audit. Routes research, extraction, critique, summarization, and adversarial review to free/cheap models. Three-tier adversarial stack (cheap/mid/top) with cross-model voices. Cost kill switch. Claude synthesizes only. Use when routing may have drifted or to set up a convergent adversarial pass.

The token-source posture for an autonomous or AFK session. LEAN’s defining move is MAXIMAL EXTERNALIZATION: push every drop of work it can onto the rest of the model stack (Groq, Cerebras, DeepSeek, Gemini, Ollama) so the scarce Claude subscription quota lasts as long as possible and does the most total work per Claude token. Builds on the `ww-keepalive` skill for the resilience mechanism when run unattended; the difference LEAN adds is the token-SOURCE posture, Claude is a precious reserve, spent only on the irreducible (final synthesis, voice/copy, irreversible judgment). Apply it as a standing per-session posture, interactive or autonomous, at any pace (one bounded task at a time, or a wide fan-out). Invoke when you say “go afk lean”, “go-afk-lean”, “qq-go-afk-lean”, “go lean”, “lean mode”, “squeeze the quota”, “stretch the main quota”, “conserve the main quota”, “push to the stack”, “max externalize”, or want a session to run while burning the minimum Claude quota.

Generate branded LinkedIn data visualizations (1200x627) using HTML/CSS and Playwright screenshots. Supports bar charts, stat grids, vertical columns, newspaper layouts, and quote+stat formats.

Two-role coordination mailbox for multi-Claude-session handoffs (build cockpit <-> overseer, or any two-or-more named roles). Append-only JSONL, atomic cursor, kind taxonomy with action hints. Load when building or debugging a coordination lane between concurrent Claude sessions.

After completing a build or reaching a milestone, steps back to ask the smart strategic questions before planning next. Data-backed, assumption-interrogating, combines options rather than picking one.

Puts your unused weekly Claude quota to work automatically. Weekend burn is a standing, every-weekend cadence (not a one-off campaign) that fires itself inside a recurring window (e.g. Friday evening through Sunday night) and burns the full weekly Claude subscription quota before it resets, finishing ONE real capability completely before starting the next so a productive weekend ships many finished things and never a pile of half-done ones. Builds on the `ww-keepalive` skill for the actual resilience mechanism (fresh `claude -p`, gates, dual-account failover) and adds the recurring-schedule + self-healing kill-switch layer on top. Invoke when you say “weekend burn”, “set up a weekend burn”, “burn the quota”, “run every weekend”, “standing weekly cadence”, “recurring autonomous campaign”, “use the whole week’s quota”, or want a repeating (not one-off) autonomous run that finishes real, categorical capabilities on a schedule instead of running once and stopping.

Use when asked to watch, babysit, audit, review, compare, or fix another agent’s work from a Codex session ID, Claude Code session/transcript, chat/thread link, PR, branch, log, or pasted run summary. Monitor until the other agent is done or blocked, reconstruct what the user asked, inspect what the agent actually changed and verified, report gaps, and optionally make scoped fixes when the user authorizes repair.

A PreToolUse hook that blocks AI-tell writing before it ever lands in a file, with a budgeted em dash rule, straight quotes, Title Case headings, banned corporate-speak, throat-clearing openers, and summary-crutch headers, plus four optional model-graded checks. Fires automatically on every Write/Edit/MultiEdit to a matching file. Nothing to invoke, nothing to remember.

“Rules for operational dashboard UX/UI: actionable error states, drill-down architecture, notification design, data visualization, progressive disclosure, performance perception, and accessibility. Load when building, reviewing, or auditing any operational or monitoring dashboard.”

Creates and edits Google Docs in your own exact house style (font, sizes, spacing, margins) via the Docs API, instead of Google’s defaults. Load this before any Google Docs work, creating a doc, writing or restyling content, converting a note into a polished doc, or building a client-facing deliverable. Covers the two reliable methods (copy a styled template vs. explicit named-style requests), hyperlinking rules, and the read-back verification step. Invoke any time a .gdoc, docs.google.com, “Google Doc”, “GDoc”, or the Docs/Drive API is in play.

Stands up a self-relaying overnight worker in about 60 seconds, a `claude -p` loop that survives 5-hour caps, session death, and reboots so a long campaign keeps running while you’re away from the keyboard. Owns the resilience mechanism (fresh `claude -p`, never `--resume`; gates + liveness checks + dual-account failover + `--permission-mode auto`; a one-paste plist install; and a standard queue.md/done.md tasklist scaffolded per campaign) so any campaign that needs to outlive a single session can build on this instead of reinventing it. Invoke when you say “keepalive”, “keep it alive”, “stand up a worker”, “self-relaying worker”, “overnight worker”, “make it survive caps”, or any autonomous-run skill needs the resilience backbone.

Working with Notion documents via the Notion MCP. Covers Notion’s custom markdown syntax, content updates, tables, callouts, highlighting, database pages, and the comments-based review workflow. Load this before ANY Notion content work.

Safe patterns for iterating Notion proposals with inline reviewer comments: verification after patches, uniqueness check before anchoring, comment-preserving rewrites. Load when iterating a reviewed Notion doc.

Autonomous overnight operation protocol for when you step away for 6+ hours with a work queue. Defines what to decide alone, what to stop for, agent launch strategy, question parking, QA standards, morning handoff format, and failure handling.

The full gauntlet for plans. Runs automatically before exiting plan mode on any multi-phase plan, rather than waiting to be invoked by name. Combines interviewing, a principles deep read, multi-domain audit, steelman, pre-mortem, convergence, and prophylactic guardrails into one comprehensive quality process.

“Mandatory end-to-end verification before claiming any feature is DONE, working, shipped, or verified. Invoke with /ww-rule15 before marking features complete, after implementing features, or when auditing existing features for cosmetic code. Triggers: claiming something works, marking DONE in plans, before commits that close features, auditing the roadmap. A feature is DONE when its intended outcome is delivered, not when its UI renders.”

Audit any agent skill, plugin, hook, or external Markdown instruction file before installing or trusting it. Fetches the SKILL.md from GitHub (or accepts a local path or URL), runs deterministic pattern scans for prompt injection, credential exfiltration, dangerous shell commands, hook installation, and supply-chain code execution, then runs a cheap-model semantic check. Returns SAFE / CAUTION / UNSAFE / UNKNOWN with specific findings. Use this skill whenever you or another skill is about to install a third-party skill, when evaluating an unknown component from `npx skills add`, when reviewing a SKILL.md from an external source, when a marketplace or repo is being added, or any time you say “audit this skill,” “is this safe to install,” “/ww-skill-auditor,” “check this skill before installing,” or point to an external skill, plugin, or hook to evaluate.

Convert straight quotes (' and ") into proper typographic quotes, apostrophes, and primes. Use this skill whenever the user asks to fix quotes, fix typography, clean up AI-generated text, polish content for publication, or apply smart quotes, and apply it proactively when producing editorial prose (blog posts, articles, marketing copy, proposals, web copy, client deliverables) so output uses proper typography from the start. Handles contractions (don't → don’t), possessives (writer's → writer’s), nested quotes, decade abbreviations ('90s → ’90s), and measurements (6'2" → 6′2″). Triggers on phrases like "fix quotes," "smart quotes," "curly quotes," "clean up this text," or whenever editing pasted content from another tool. Skip inside code, file paths, regex, and technical syntax where literal quote characters matter.