protoLabsAI

Analyze A/B test results with statistical significance, sample size validation, confidence intervals, and ship/extend/stop recommendations. Use when evaluating experiment results, checking if a test reached significance, interpreting split test data, or deciding whether to ship a variant.

Analyze and prioritize a list of feature requests by theme, strategic alignment, impact, effort, and risk. Use when reviewing customer feature requests, triaging a backlog, or making prioritization decisions.

Generate an Ansoff Matrix analysis mapping growth strategies across market penetration, market development, product development, and diversification. Use when considering growth options, planning market expansion, or evaluating strategic growth paths.

Identify the first beachhead market segment for a product launch. Evaluates segments against burning pain, willingness to pay, winnable market share, and referral potential. Use when choosing a first market, targeting an initial customer segment, or planning market entry strategy.

Brainstorm product/feature ideas with multi-perspective ideation (PM, Designer, Engineer), then prioritize the best five. Works for an existing product (continuous discovery) or a new product (initial discovery). Use when generating feature ideas, ideating solutions for an opportunity, or running a product-trio ideation session.

Brainstorm team-level OKRs aligned with company objectives — qualitative objectives with measurable key results. Use when setting quarterly OKRs, aligning team goals with company strategy, drafting objectives, or learning how to write effective OKRs.

Generate a Business Model Canvas with all 9 building blocks. Use when creating a business model, documenting how a business creates value, or analyzing an existing business model.

Create sales-ready competitive battlecards comparing your product against a specific competitor — positioning, feature comparison, objection handling, and win/loss patterns. Use when preparing sales teams, creating competitive materials, or responding to 'why not competitor X?'

Full clean rebuild of ORBIS — wipes web/dist, dist-sdist, src-tauri bundle, staged sidecar binary, pyapp env cache, WKWebView state, sidecar.log, and stale unix sockets, then rebuilds frontend → sdist → pyapp sidecar → tauri bundle in the correct order, and launches with stderr captured. Use when voice doesn't work, the orb shows weird state, "Load failed" appears anywhere in the UI, or anything feels off after a code change.

Step-by-step workflow for adding a new processing component to the ORBIS agent pipeline

Step-by-step workflow for integrating a new stage into the ORBIS audio/processing pipeline

Integrate a new ML component (classifier, reranker, etc.) into an existing agent pipeline by reading the codebase, identifying hook points, implementing the module, and wiring it in.

Commit pending changes, bump the patch version tag, and publish a beta release

Scaffold and plan a new marketing site — branding audit, stack selection, hero design, and deployment config

Integrate a new ML component (classifier, reranker, embedder, etc.) into an existing pipeline by locating insertion points, implementing the module, and wiring it in.

Wire a new FrameProcessor into ORBIS's pipecat pipeline at the correct hook point

Use this when asked to build, create, write, scaffold, or publish a protoAgent plugin — e.g. "make a plugin that …", "add a plugin for X", "package this as a plugin", "write a plugin that adds a tool/dashboard/workflow", "publish a plugin so others can install it". Covers the plugin contract (manifest + register()), the full contribution surface (tools, subagents, SKILL.md skills, workflows, console views, routes, MCP servers, config/secrets/settings), the conventional repo layout, testing, and distribution by git URL — with the safety model. Not for: using an already-installed plugin, or extending via a one-off SKILL.md skill or MCP server (smaller asks — see the Skills / MCP guides).

/due-diligence — validate a technology or architecture choice with the due-diligence workflow: codebase map + external research in parallel, antagonist + claim verification, and a cited adopt/build/defer verdict.

Use this when writing, drafting, or updating an Architecture Decision Record (ADR) or a plan document for this repo — e.g. "write an ADR for X", "record this decision", "author the ADR for this design", "draft the plan doc". Covers the house MADR shape, numbering, the index row, the docs-nav regeneration step, and the VitePress traps that fail the docs build.

/code-review — adversarial PR review via the code-review workflow: four parallel finder angles, dedup, an evidence-checking verify pass, and a findings report. Falls back to a two-axis inline review when the workflow isn't available.

When the user wants to SEE, render, visualize, preview, or "show me" a chart, diagram, mock-up, table, or interactive widget — render it with the show_artifact tool instead of writing files to disk.

/grill — a relentless interview that sharpens a plan or design before any of it is built.

/standup — an operational status report on everything this agent currently owns: tasks, goals, schedules, background work, and what needs the operator's call.

/writing-skills — the house discipline for authoring SKILL.md skills that behave predictably. Loads the rules; the skill_writer subagent applies them.

Use whenever anything of value sits behind a post — a paid sponsorship, a gifted product, an affiliate link, an employee writing about their own employer, or the brand's own product. Covers what the FTC actually requires, where the disclosure has to sit, and the wording that doesn't count. Triggers: "we're doing a paid partnership", "they sent us the product", "affiliate", "our employees are posting about it", "do we need to say it's an ad", "is this compliant".

Use the moment scheduled content becomes a liability — a crisis at the company, an outage, a public tragedy, a story breaking about the brand or its industry, a post going wrong. Covers holding the queue, drafting a holding statement, and deciding when to resume. Triggers: "hold everything", "pause the posts", "we have a situation", "this is blowing up", "should we still post today", "something happened".

Use to close the loop — record what published posts actually did, and work out what that means for the next calendar. Run it monthly, or before planning. Triggers: "how did last month do", "what's working", "here are the numbers", "review performance", "should we keep doing X", "what should we do more of".

Use to write social copy — one post or a batch, for any platform. Covers hooks, native structure per surface, the lint-and-fix loop, and getting to an approved queue item. Triggers: "write a post about X", "draft something for LinkedIn", "turn this into a tweet", "we need a post for the launch", "draft the ideas in the queue".

Use to research and record how a platform currently behaves — the length band that performs, hashtag counts, where the fold falls, whether links get demoted, what flops. Run it when social_check reports "no norms on file", when norms are flagged stale, or when the operator says results have changed. Triggers: "research the norms", "what works on LinkedIn now", "our reach dropped", "refresh the platform rules", "no norms on file".

Use to build or overhaul the brand kit — the file that makes every draft sound like this brand instead of generic marketing. Run it before the first post, and again when the positioning, audience, or voice changes. Triggers: "set up the brand", "who are we talking to", "define our voice", "the posts don't sound like us", "onboard me", or any drafting request made when social_brand_kit reports no kit exists.

Use to plan what goes out and when — a two-week or monthly calendar balanced across the brand's content pillars and platforms, seeded with concrete post ideas. Triggers: "plan our content", "what should we post", "build a calendar", "we need a month of posts", "fill the queue", or a review of what's already scheduled.

Use to prepare the half of social that isn't posting — replies to comments, responses to criticism, DM openers, and the accounts worth engaging with. Triggers: "help me reply to this", "someone commented X", "how do I respond to this criticism", "draft replies", "what should I engage with", "we got a complaint on LinkedIn".

Produce or modify Word documents (.docx) as real files — reports, memos, letters, proposals, or any deliverable the operator will open in Word. Use when the requested output is a Word file, or when an existing .docx needs content extracted, edited, or restructured. Not for spreadsheets, decks, or PDFs — those have their own skills.

Read, extract from, combine, split, watermark, fill, or produce PDF files. Use whenever a .pdf is the input or the deliverable — merging reports, pulling text or tables out, rotating scans, filling forms, or generating a polished PDF from content.

Build or edit slide decks (.pptx) as real files — pitch decks, briefings, status readouts — or extract content from existing ones. Use whenever slides or presentations are the input or the deliverable. Not for documents, spreadsheets, or PDFs.

Produce, edit, or clean spreadsheets (.xlsx/.csv/.tsv) as real files — reconciliations, trackers, budgets, data cleanups, format conversions. Use whenever the deliverable is a spreadsheet or an existing one needs work, including messy tabular data that needs restructuring. Not for Word/PDF/deck outputs.

Compose the operator's daily brief — schedule, things waiting on them, ongoing work, and heads-ups — rendered as a polished HTML artifact, optionally on a recurring schedule. Run only when the operator explicitly asks for their brief (or /daily-brief), or when executing a scheduled brief task. An ordinary question about today's calendar is answered directly, not with a brief.

Conversational first-run companion for the Cowork archetype — verify the workspace is ready, import existing Claude Code/Cowork state, prove the setup on a real task, learn the operator's writing voice, and wire the first scheduled task. The mechanical setup (archetype, plugins, dependencies, folders) belongs to the setup wizard; this skill picks up where it leaves off.

Reflective maintenance pass over long-term memory — merge duplicates, retire stale or completed items, sharpen what stays. Use when the operator asks to clean up/organize memory, when recall keeps surfacing outdated facts, or periodically as a scheduled task.

Turn work into a scheduled task that runs unattended — "every morning", "each Friday", "remind me in an hour", "run this weekly" — or adjust an existing schedule. Use whenever the operator wants something recurring, delayed, or re-timed.

Answer any question about protoCLI usage, features, configuration, and troubleshooting by referencing the official user documentation. Also helps users view or modify their settings.json. Invoke with `/proto-helper` followed by a question, e.g. `/proto-helper how do I configure MCP servers?` or `/proto-helper change approval mode to yolo`.

Review local code changes with git diff and update the official docs under docs/ to match. Use when the user asks to document current uncommitted work, sync docs with local changes, update docs after a feature or refactor, or when phrases like "git diff", "local changes", "update docs", or "official docs" appear.

Use browser automation to interact with web pages - navigate, click, fill forms, take screenshots, and extract content from websites. Perfect for testing, data collection, and web interactions.

Reference guide for all agent harness safety features — doom loop detection, scope lock, git checkpoints, observation masking, sprint contract, reminders, repo map, behavior verification, and multi-sample retry

Review changed code for correctness, security, code quality, and performance. Use when the user asks to review code changes, a PR, or specific files. Invoke with `/review`, `/review <pr-number>`, or `/review <file-path>`.

Automates terminal UI screenshot testing for CLI commands. Applies when reviewing PRs that affect CLI output, testing slash commands (/about, /context, /auth, /export), generating visual documentation, or when 'terminal screenshot', 'CLI test', 'visual test', or 'terminal-capture' is mentioned.

Audit the repository's docs/ content against the current codebase, find missing, incorrect, or stale documentation, and refresh the affected pages. Use when the user asks to review docs coverage, find outdated docs, compare docs with the current repo, or fix documentation drift across features, settings, tools, or integrations.

Reviews pull requests with code analysis and terminal smoke testing. Applies when examining code changes, running CLI tests, or when 'PR review', 'code review', 'terminal screenshot', 'visual test' is mentioned.

Add new content to an existing world pack. Batch add monsters, NPCs, items, locations, and encounters.

Enter focused combat mode with only combat-relevant tools available. Optimized for fast, decisive tactical play.

Enter focused exploration mode with tools for session management, time tracking, scene framing, encounters, leads, and relationships.

Generate print-ready books from an existing world pack. Creates GM book, player book, bestiary, and campaign setting.

Start or resume an RPG adventure. Creates worlds, characters, and runs the game.

Enter focused session-management mode with session, player, turn coordination, and time tools. Optimized for setting up and managing game sessions before or between play.

Enter focused world-building mode with only generation, world pack, and expansion tools available. Optimized for creating complete RPG worlds from scratch or expanding existing ones.

Codex-native operational orchestration for protoLabs Studio. Use when the user wants autonomous triage, backlog supervision, board operations, agent coordination, or multi-step operational decision-making across the portfolio.

Codex-native deep work mode for protoLabs Studio. Use when the user wants autonomous backlog processing, PR flow supervision, board grooming, and sustained operational work until the system is quiet.

Codex-native analysis workflow for root-cause investigation. Use for bugs, performance issues, architectural confusion, and other cases where the main deliverable is findings, not code.

Codex-native codebase research workflow for protoLabs Studio. Use before planning or implementing a feature when you need to map existing patterns, constraints, related files, and integration points.

Codex-native validation workflow for technology choices and architectural approaches. Use when the user wants evidence-based comparison, external validation, tradeoff analysis, or risk assessment before committing to a path.

Codex-native full project lifecycle workflow for protoLabs Studio. Use when the user wants to go from project idea to PRD, milestones, board features, and launch readiness.

Use when the user shares a job posting (URL or text) or asks to evaluate fit, tailor a CV/resume, or write a cover letter — "should I apply to this", "evaluate this job", "tailor my CV for X", "write a cover letter for the role at Y", "apply to this". The execution half of the Career Coach: score fit against the rubric, then draft tailored, honest, verified documents. For interview practice see interview-coach; for positioning / offers / negotiation see career-strategy; for skill gaps see upskill.

Use when the user wants the FULL, filed application worked up for a specific role — "build the role packet", "do the whole application for this job", "tailor everything for the Merck AD role", or when they want a resume + evidence map + ATS skills list + cover letter produced as saved files they can review and edit. A gated, human-approved pipeline that files a folder per role and produces each artifact one confirmed step at a time. For a quick fit read or a single document, use job-application-assistant instead; for the fast autonomous version, run the `apply` workflow.

Writing-quality pass for application materials — make a draft read like a sharp human wrote it, not a template or an LLM. Use on any resume bullet, cover letter, or message before it ships. Complements the coach's writing-style discipline with a final de-slop edit.

Use for the coaching conversations that aren't about a single application — positioning and personal brand, deciding what roles to target, weighing or comparing offers, salary and benefits negotiation, whether to take a job, and career-direction / reflection. Triggers: "help me think about my career", "what roles should I target", "should I take this offer", "how do I negotiate this salary", "compare these two offers", "how do I position myself".

Use to prepare for or practice interviews — "help me prep for the interview at X", "run a mock interview", "practice behavioral questions", "how do I answer 'what's your weakness'", "what should I ask them". Builds a STAR answer bank, runs realistic mock interviews with honest per-answer feedback, and prepares questions to ask the interviewer.

Use to turn the jobs someone is chasing into a concrete learning plan — "what should I learn", "where are my skill gaps", "how do I become competitive for X roles", "make me a study plan". Aggregates the tracked pipeline (or a single posting) against the profile, builds a prioritized gap heatmap, and produces a learning plan with real, web-searched resources.

Use this whenever the operator wants to test a trading idea on history — e.g. "backtest a 20/50 moving-average cross on SPY", "does RSI mean-reversion work on AAPL", "how would a breakout strategy have done on BTC". Drives a specify → backtest → benchmark → check-robustness → honest verdict loop.

Use this when the operator asks whether a factor / signal predicts returns — e.g. "does momentum work", "is low-vol still an edge", "rank these factors", "what's working in the market right now". Drives an evaluate → read-IC → context → honest verdict loop over the Alpha Zoo.

Use this ONLY when the operator explicitly asks to place, size, or manage a simulated/paper order — e.g. "buy 10 shares of AAPL in the paper account", "paper-trade this setup", "check my paper positions", "what's my paper P&L". Drives a confirm-armed → size → preview → operator-approval → fill loop. Never initiates a trade on your own.

Use this whenever the operator asks you to research, analyze, or give a read on a specific stock, ETF, or crypto — e.g. "what's the setup on NVDA", "should I be watching SPY", "analyze BTC", "bull/bear case for AAPL". Drives a quote → trend → fundamentals/context → news → structured read with risks.

Use this when the operator shares a trade journal / broker export and wants to understand their trading behavior — e.g. "review my trades", "what am I doing wrong", "analyze my journal", "why am I losing money even with a high win rate". Drives an ingest → profile → name-the-biases → Shadow-Account → fixes loop.

The OODA strategist loop for the SpaceTraders fleet — the agentic BRAIN that steers the deterministic autopilot engine. The engine now self-stabilizes (capital-disciplined buys, confirmed-fresh routing, and a position-sticky plan that holds routes + stations probes), so the strategist steers POLICY SLOWLY — daily or on a real regime change — not every window. Each tick: Observe (st_report) → Orient (recall lessons, compare vs the goal) → Decide → Act (st_strategy / st_tune — one bounded change) → Learn (memory_ingest). Triggers on "manage the fleet", "run the fleet to a million", "strategize the fleet", "keep steering the fleet".

Use to run the SpaceTraders fleet autonomously toward a goal — default: make the most credits per hour. Drives the fleet-commander + autopilot engine, and sets up the recurring loop (scheduler tick + goal). Triggers on "manage the fleet", "make money autonomously", "run the fleet", "maximize credits", "keep the fleet busy".

Use when the operator wants to play SpaceTraders — command the fleet, check the agent, mine asteroids, trade goods, or work a contract in the live galaxy (spacetraders.io). Drives the loop: status → scan → navigate → mine/trade → sell → contract, using the st_* tools. Triggers on "check our agent", "what ships do we have", "go mine", "find a market", "fly to", "work the contract".

Use when working a SpaceTraders PROCUREMENT contract — "deliver N units of X to waypoint Y". Covers the efficient buy-low / deliver pattern and when to delegate to the procurement-run workflow vs the trader/navigator crew. Triggers on "work the contract", "fulfill the contract", "we need to deliver", "get the ore".

Use FIRST, when pointed at a repo this team hasn't worked before (or to re-check a repo's readiness) — BEFORE decomposing or dispatching any feature. Scans the repo for the preconditions a coding-agent loop needs, AUTO-FIXES the safe/deterministic gaps (beads workspace, ignored agent scratch, the build/test gate) by delegating to the coder, and BOARDS the judgment gaps (a project grounding doc, PR CI) as features. Ends with a readiness report + a human gate. Do NOT use to plan features or write code — that's `decompose-project` and the coder.

Use to run a retro on the coding loop's own track record and distill the lessons into durable grounding. Mines the board's attempt/outcome history (board_retro) for RECURRING failure classes + flow stats, then (a) WRITES each new lesson to the knowledge graph (domain "loop-lessons") — which the coding loop injects into coder prompts at build time — and (b) PROPOSES the same gotchas for PROTO.md in a dated retro report for a human to apply. The self-improving half of the flywheel. Runs unsupervised on a schedule (the dream/distill pattern) — it PROPOSES, it does not auto-edit PROTO.md or commit; a human applies the report. Does NOT write code.

Use when asked to plan, decompose, or spec out a new project/idea before any code is written. Runs the adversarial decomposition pipeline (idea → outline → MADR decision records → epics › milestones › features with EARS acceptance criteria), writing the plan as docs in the working tree AND populating the beads board — with an antagonist hardening every step and a human approval gate per epic. The board's loop then builds the `ready` features. Do NOT use to write code; this only plans.