← Documentation

Documentation · August 3, 2026

Skill Hub

What a skill is, the full audited catalogue with its Neural Credit price and minimum plan, how to grant a skill to a named OmniForce, and what each refusal means.

A skill is one audited capability an OmniForce can use: booking an appointment, drafting a quote, answering from your own documents, asking a customer for a review. An OmniForce without skills can hold a conversation. An OmniForce with skills will be able to act.

The Skill Hub is where you decide which of those capabilities each Force has. You reach it at eyeclaw.app under Skills, or from the Tools tab of the EyeClaw cockpit inside the desktop OS. Both surfaces grant through the same rules, so it does not matter which you use.

Where skills stand today

Skill execution is not available yet, and nothing on this page is charged.

What works today is the entitlement half: you can grant a skill to a named OmniForce, the platform enforces the audit, risk, plan and credit gate at the moment you grant it, the grant is recorded against that Force, and you can revoke it. That record is what a Force will run from.

What does not exist yet is the executor — the part that takes a granted skill and performs the work. Because nothing runs, no Neural Credits are spent on skills at all: not when you grant one, not on a Heartbeat beat, not anywhere. The prices below are real and they are what one run will cost once execution ships; they are not being billed to you now.

Everything else in this article — the catalogue, the gate, the refusal reasons — describes what is live today. Where a price is quoted, read it as what a run will cost, not as a charge you are incurring.

What every skill carries

Each skill in the catalogue has four properties, and all four are enforced before it can be granted:

  1. An audit status. Only skills that have passed Eyeleven's audit are usable. A skill still in the pipeline shows In review and cannot be granted by anyone, on any plan.
  2. A risk level — LOW, MEDIUM, HIGH, or CRITICAL.
  3. A price in Neural Credits, per run. A skill listed at 3 credits will cost 3 credits every time it runs once execution ships — per run, not once when you grant it. Granting has never cost credits, and today running does not happen at all.
  4. A minimum plan. Some skills require Nexus, Hyperion, or Omnis.

Skills are granted to a Force, not to your workspace

This is the part that most often surprises people, so it is worth stating directly.

A skill is granted to one named OmniForce. It is not switched on for the whole workspace. If you have a Sales Force and a Support Force, granting Generate Quote to the Sales Force does nothing for the Support Force, and vice versa.

That is deliberate. It lets you give each Force exactly the capabilities its job needs and no more — your booking Force can book and remind, without also being able to send quotes or reactivate dormant customers. A skill does nothing at all until some Force holds it.

The Skill Hub shows every Force in your workspace along with how many skills each one currently holds, and each skill card shows which Forces already have it.

Granting and revoking

To grant a skill:

  1. Open the Skill Hub and find the skill. You can search by name or outcome, and filter by category, by risk level, or by Available to me.
  2. On the skill's card, under Granted to, select the Force you want to give it to.
  3. The Force now holds the skill. The grant is recorded and audited immediately; nothing starts running and nothing is charged, because execution has not shipped yet.

To revoke, select the same Force again. Revoking is always permitted — taking a capability away never requires an entitlement, so you can remove a grant even if the skill has since become unavailable to grant fresh. A skill that can no longer run is exactly the one you most want to be able to remove.

Two things that follow from how this works:

  • A grant is not a bypass. The same checks are designed to run again every single time the skill executes. If your plan changes or your credits run out, a granted skill will simply not run until that is resolved.
  • If the change is refused, nothing is recorded. The Hub rolls the control back to what the runtime actually holds and shows you the reason, rather than leaving you believing a grant succeeded when it did not.

The catalogue

Twenty skills ship today. These seventeen can be granted, subject to your plan and your credit balance.

The Credits per run column is the price of one run once skill execution ships. No skill is billed today — see Where skills stand today.

Skill What it does for your business Credits per run (once execution ships) Minimum plan
Book Appointment Creates a calendar booking and confirms it with the customer. 5 Ascent
Appointment Reminder Messages the customer before their appointment and confirms they are coming, so fewer people simply do not turn up. 2 Ascent
No-show Recovery Reaches out after a missed appointment and offers to rebook, turning a lost slot into a booking. 3 Nexus
Browse Catalogue Walks the customer through your services or products and their prices, so nobody waits for an answer on cost. 1 Ascent
Generate Quote Drafts a priced quote from your catalogue and what the customer asked for. 5 Nexus
Handle Objection Answers a price or trust objection using your own positioning, keeping a hesitant buyer in the conversation. 2 Nexus
Qualify Lead Scores and qualifies an inbound lead from the conversation so far. 3 Ascent
Ask for a Review Asks a happy customer for a public review at the moment people actually say yes. 2 Ascent
Post-visit Check-in Asks how the visit went and flags anyone unhappy for a human before they post about it. 2 Ascent
Reactivate Dormant Customer Opens a conversation with a customer who has not been in touch for a while. 3 Nexus
Escalate to a Human Hands the conversation to a person, with a summary of what happened so far. 1 Ascent
Match the Customer's Language Detects the customer's language and continues in it, so one number serves Spanish, English, and Portuguese. 1 Ascent
Answer from Knowledge Retrieves from your Business Memory files and answers the customer. 1 Ascent
Read a PDF Extracts text and tables from a PDF, so you can upload a price list or brochure and have the Force quote from it. 2 Ascent
Read a Spreadsheet Reads an XLSX or CSV of services, prices, or stock so the Force can use it. 2 Ascent
Read a Word Document Reads a DOCX of policies, scripts, or terms, so they are answered word for word rather than improvised. 2 Ascent
Start Follow-up Sequence Enrols the customer into a nurturing or follow-up sequence. 3 Nexus

Three more exist in the catalogue but cannot be granted by any workspace today, whatever plan you are on:

Skill Credits per run (once execution ships) Why it is not available
Bulk CRM Update 8 Rated HIGH risk. Eyeleven does not release skills at that level to customer workspaces yet. No plan and no credit balance unlocks it.
Send Payment Link 5 Still in Eyeleven's audit, so it shows as In review. It is also rated HIGH risk.
Charge Customer 10 Still in Eyeleven's audit, so it shows as In review. It is also rated CRITICAL, the level that is never released.

They are listed in the product rather than hidden, so you can see the full shape of the catalogue and what is coming.

The Documents skills are built on open-source work

Read a PDF, Read a Spreadsheet, and Read a Word Document are built on Anthropic's open-source skills, published under the MIT licence. They are not Eyeleven's own work, and the product says so: each of those three cards shows the upstream project it came from, the licence, and a link to the source.

Every other skill in the catalogue is Eyeleven's own. The reason the whole catalogue is worth paying for is not that the text behind a skill is scarce — it is that each one is wired into your WhatsApp, your CRM, your calendar, and your Knowledge Base, audited before release, metered, and revocable. Eyeleven does not list imported work without crediting it.

What each refusal means

When a skill cannot be granted, the Hub says which of four things is stopping it. They are checked in this order, so the first one that applies is the one you see.

In review (or Draft, Deprecated, Rejected) — the skill has not passed Eyeleven's audit, so no workspace can use it. Nothing you do changes this; it clears when the audit finishes, if it passes.

Held back by risk policy — the skill is rated HIGH or CRITICAL, and Eyeleven does not release skills at those levels to customer workspaces. No plan and no credit balance unlocks it. This is a platform decision, not an entitlement you can buy.

Needs the [plan] plan — your workspace is on a lower plan than the skill requires. Upgrading makes it grantable and changes nothing else about the skill. You change plans in the Billing app in the desktop OS; see Plans & Neural Credits.

Not enough Neural Credits — one run would cost more than your current balance. The Hub tells you both numbers. Top up with a Neural Credit Pack, or wait for your cycle to reset, and it becomes grantable again.

The first two cannot be resolved from your side. The last two can, and the Hub marks them differently for exactly that reason.

How credits will be spent

Today: they are not. No skill has ever been executed by the platform, so no skill has been charged. Granting, revoking, and Heartbeat beats are all free, and your balance is untouched by anything on this page.

Once skill execution ships, this is the contract:

  • A skill's cost is charged when a run completes — after it has passed the audit, risk, plan, and credit checks and the work has actually been done.
  • A skill refused by any of those checks is never charged. A blocked run costs you nothing.
  • The cost is per run, every run. Granting a skill costs nothing on its own — that will not change.
  • If your balance falls below a skill's cost, that skill will not run until you top up or your billing cycle resets. Other, cheaper skills keep working in the meantime.

Your balance is shown in the Skill Hub, in the EyeClaw cockpit, and in the Billing app.

Tips

  • Grant narrowly. Give each Force the skills its role needs and nothing else. It is easier to add one later than to work out why a Force did something you did not expect.
  • Watch the price, not just the plan. Answer from Knowledge is priced at 1 credit and Book Appointment at 5. Once execution ships, a Force that answers questions all day and books occasionally will spend very differently from one doing the reverse.
  • Load your documents first. The three Documents skills are what turn a price list or a policy document into something a Force can quote from accurately. See Quantum Files & Knowledge Base.
  • Revoke without hesitation. Revoking is instant, always allowed, and reversible. It is the fastest way to stop a behaviour you did not intend.

Practical check

A grant is not execution

Skill permissions describe what an OmniForce may be allowed to use. The current skill executor has not shipped; configuring a grant does not run the skill.

  • Inspect which role and workspace receive the grant.
  • Keep sensitive actions behind the required human review.
  • Do not treat a configured skill as a completed task or a credit charge.
Read current product scope ↗