Productivity

Make, adapt, and revise videos with Hypit from references or briefs, including SVML/SVS/SVRun authoring, project components, and Runtime or credential setup.

Builds a unit-economics and cohort .xlsx with live formulas across separate assumptions, cohort, and output tabs plus checks, computing CAC, LTV, LTV/CAC, and CAC payback from a monthly cohort retention grid. Use when someone needs unit economics, an LTV/CAC analysis, a payback model, or a cohort retention build. Do not trigger for editing an existing workbook through the Claude for Excel add-in.

Builds a Base/Bull/Bear scenario switching layer for a model in a .xlsx with a live selector, CHOOSE or INDEX assumption links, a side-by-side output comparison, and a recalc check, then delivers the workbook. Use when someone wants an auditable scenario toggle instead of Excel's hidden Scenario Manager. Do not trigger for editing an existing workbook through the Claude for Excel add-in.

Builds one- and two-variable sensitivity (Data Table) analysis on a model output in a .xlsx with live formulas and a recalc check, then delivers the workbook. Use when someone wants to see how an output (NPV, IRR, EPS, margin) moves as one or two inputs change. Do not trigger for editing an existing workbook through the Claude for Excel add-in.

Ranks variable sensitivities to a key output and builds a tornado chart by swinging each input over a defined range, computing output impact, and producing a horizontal bar chart data structure.

Enforces the three-tab separation principle for any Excel model, producing a full design for a dedicated Inputs tab, one or more Calculations tabs, and a clean Outputs tab with a conventions checklist.

Builds a driver-based revenue forecast .xlsx with live formulas across separate drivers, build, and output tabs plus checks, rolling a customer or units schedule (beginning, adds, churn, ending) into revenue with Base/Bull/Bear scenario toggles. Use when someone needs a bottom-up revenue model, an ARR build, or a driver-based forecast for planning or a board deck. Do not trigger for editing an existing workbook through the Claude for Excel add-in.

Helps you answer sales objections (price, timing, "send me info", competitors, stalls) using rules weighted by how many of the 20 most-watched videos on the topic repeat each one, and flags where the experts disagree so you can A/B test instead of guessing. Use when the user says "handle this objection", "they said it's too expensive", "they want to think about it", "they asked me to send info", "they're going with a cheaper competitor", "the prospect is stalling", or "how do I respond to this pushback".

Writes and plans LinkedIn outreach (connection requests, DMs and prospecting) using rules weighted by how many of the 20 most-watched videos on the topic repeat each one, and flags where the experts disagree so you can A/B test instead of guessing. Use when the user says "write a LinkedIn DM", "LinkedIn connection request", "LinkedIn prospecting", "cold outreach on LinkedIn", "build a LinkedIn lead list" or "review my LinkedIn message".

Plans, runs and reviews sales discovery calls (questions, structure, qualifying, uncovering pain) using rules weighted by how many of the 20 most-watched videos repeat each one, and flags where experts disagree so you can A/B test instead of guessing. Use when the user says "prep my discovery call", "what should I ask on a sales call", "qualify this prospect", "review my discovery call", "structure a first sales call" or "how do I uncover pain".

Write and review cold emails using what the 20 most-watched cold email videos of 2026 actually agree on, weighted by how many of them repeat each rule, and flag the choices they disagree on so you A/B test those instead of trusting one guru. Trigger on "write a cold email", "review my cold email", "cold email sequence", or /cold-email.

Builds B2B lead magnets and inbound lead generation plans using rules weighted by how many of the 20 most-watched videos on the topic repeat each one, and flags where the experts disagree so you A/B test instead of guessing. Use when the user says "create a lead magnet", "B2B lead magnet ideas", "get more inbound leads", "build a quiz funnel", "lead magnet follow-up emails", or "where should my leads come from".

Builds B2B positioning and messaging (value proposition, differentiation, how to describe what you sell) from rules weighted by how many of the 20 most-watched videos on the topic repeat each one, and flags where experts disagree so you A/B test instead of guessing. Use when the user says "position my product", "write our value proposition", "how do we stand out from competitors", "fix our messaging", "describe what we sell", or "B2B positioning".

Plans, writes, and reviews LinkedIn posts, hooks, profiles, and engagement routines to grow an audience and turn followers into clients, weighting each rule by how many of the 20 most-watched videos on the topic repeat it and flagging where experts disagree so the user can A/B test instead of guessing. Use when the user says "write a LinkedIn post", "grow on LinkedIn", "fix my LinkedIn hook", "LinkedIn content plan", "optimize my LinkedIn profile", "get clients from LinkedIn", or "why is my LinkedIn post not getting reach".

Helps price a B2B product or service (strategy, packaging, raising prices, value-based pricing) using rules weighted by how many of the 20 most-watched videos repeat each one, and flags where the experts disagree so you A/B test instead of guessing. Use when the user says "help me price this", "should I raise my prices", "value-based pricing", "how should I package my tiers", "stop charging hourly", "hybrid or usage pricing", or "review my pricing page".

Builds cold call openers, scripts, first-30-second handling, and meeting-booking asks from rules weighted by how many of the 20 most-watched cold calling videos repeat each one, and flags the choices where experts disagree so you A/B test them instead of guessing. Use when the user says "write my cold call script", "help me with my cold call opener", "how do I handle the first 30 seconds", "book more meetings by phone", "cold calling objections", or "fix my cold calls".

Writes sales follow-ups and closing next steps using the rules repeated most often across the 20 most-watched videos on the topic, weighting each rule by how many of those videos say it, and flags any expert disagreements to A/B test. Use when the user says "write my follow-up", "the prospect went quiet", "how do I close this deal", "what do I send after the sales call", "help me book the next meeting", or "they said they need to think about it".

>-

Read-only review of Caveman Cloud evidence: cost, Cave Score, workflows, traces, latency, errors, routing, savings. Use when asked what Caveman found or where LLM spend goes.

Read-only repository explorer for cold-start orientation, broad cross-file localization, or when a direct search failed. Skip it when the exact file or symbol is already named. Returns path:line citations only; its reads stay out of main context.

Find and label every LLM workflow in the repository so Caveman Cloud groups spend by workflow instead of one bucket. Use for "discover workflows" or breaking LLM spend down by workflow.

Inspect Caveman Cloud's experiment lifecycle and block unsafe execution. Use when asked to start, approve, cancel, promote or roll back a Caveman experiment.

Act on a Caveman learn report - review the ranked token sinks, apply cost-lowering fixes with per-edit consent, and report what those fixes returned. Use when asked to lower an agent's token cost, what caveman has saved, to trim a heavy CLAUDE.md, or to offload re-pasted context into cavemem.

Hand the current conversation off to a fresh background agent that picks up the work immediately.