convex-cost Skill for Claude Code: What It Does and How to Install It

Makes Claude rank the Convex functions that drive your spend, show how they grow and name the cheapest fix for each.

Installs: 38.3K; on skills.sh
Stars: 68; on GitHub
Updated: Oct 2026; last repo commit

This skill makes Claude find what drives the cost of a Convex app. It reads usage evidence from the deployment, ranks functions by data read per call times call volume and shows how each top driver grows as traffic and tables grow. For each one it names the cheapest fix, such as an index instead of a scan, take or paginate instead of collect, an aggregate component or a cached read.

SKILL.md asks for relative guidance and growth shapes, not made-up dollar totals, and points to the pricing page for absolute numbers. The usage evidence comes from the Convex MCP tools insights, tables and functionSpec, and SKILL.md says insights works on cloud deployments with a signed-in user, not on previews. Its folder holds only SKILL.md.

Good for

Convex developers who want to know which functions cost the most
Teams checking how a table that keeps growing will affect their bill

Install convex-cost

Claude Code

Copy this prompt.

Install the convex-cost skill in Claude Code. Run this command:

npx skills add https://github.com/convex/agent-skills --skill convex-cost --skill convex-deploy-guard

Then check that it loaded and tell me how to use it.

2
Paste it into Claude Code. It shows the change and asks before running it.

Requires Node.js.

SKILL.md hands work to convex-deploy-guard, so the first command installs it too. It also mentions convex-advisor and convex-expert as optional hand-offs; install those separately if you want them. Convex's docs say the skills CLI installs skills into .agents/skills/ in your project. The "All Convex skills" command installs every skill in the repo, which has 33 skill folders. One of them, convex-improve-convex-plugin, is listed in the Convex docs as sending your coding session's transcript to the Convex team. Run this skill with /convex-cost in Claude Code or $convex-cost in Codex, or let the agent pick it when the task fits. This skill reads your deployment through the official Convex MCP server, which is not part of the skills. Add the server command to your client's MCP settings; the Convex docs also describe an official plugin that starts it for you. This skill works on your project folder, so this page offers no Claude app upload route.

Built from the package data we hold. Not yet tested by us.

What's in the skill

When it loads

Preview Convex spend — rank functions by bytes/documents-read × call-volume from insights, project each cost driver's growth curve, name the cheapest fix; confirm-cost for paid actions.

Sections in SKILL.md

Workflow
Rules

Key facts

Author: Convex (GitHub organisation convex; the vendor's own repo)
License: Apache-2.0 (stated in the README and the LICENSE file)
Installs on skills.sh: 38,317 (Nov 2, 2026)
GitHub stars: 68 for the whole convex/agent-skills repo (Nov 2, 2026)
Last repo commit: Oct 29, 2026
Audits on skills.sh: Gen Agent Trust Hub: pass, Socket: pass, Snyk: pass
What it includes: 1 file: SKILL.md (the repo tree lists nothing else in the folder)
Part of: The convex/agent-skills repo: 33 skill folders in the tree. The README names 6 skills, and this one is not among them

What it includes

Uses deploy-guard to classify and announce the deployment before reading
Read · Guard the target
Reads bytes read with insights, row counts with tables, functions with functionSpec
With no traffic yet, it estimates from the shapes of your queries instead.
Network · Gather spend evidence
Ranks functions by data read per call times call volume
SKILL.md says a cheap function called constantly can cost more than an expensive rare one.
Attribute the cost
Shows how each top driver scales, for example a full-table collect grows with the table
An indexed take(n) stays flat. SKILL.md asks for the shape of the curve, not a false-precision dollar figure.
Project growth
Suggests an index, take or paginate, an aggregate component or caching for each driver
It records a cost finding that points to convex-expert and convex-advisor for the actual change.
Try: “Which of my Convex functions will cost the most as we grow?”
Name the cheapest fix
States the price and recurrence and waits for an explicit yes before anything metered
SKILL.md gives a domain purchase, cloud provisioning and a plan change as examples.
Confirm cost for paid actions
Lists drivers with evidence, growth shape and fix, plus a plain bottom line
Absolute prices are left to the pricing page.
Report

From the official docs

Permissions

Low risk

What it can do

Read your data
Create and change things
Delete things
Read-only modeHow to turn it on

Watch out for

Paid actions need your explicit yes
SKILL.md says that before anything metered, such as a domain purchase or a plan change, Claude states the price and recurrence and waits for a yes. github.com
The Convex MCP server limits production access by default
The Convex docs say the MCP server is safe by default: in production deployments agents cannot access PII and can only perform read-only operations. Write access and PII reads on production need --dangerously-enable-production-deployments or --cautiously-allow-production-pii. docs.convex.dev

Security advisories

No published security advisories as of Nov 2, 2026.

Technical details

Turning on read-only	SKILL.md describes the cost read as read-only over dev and prod and says insights is cloud and user-auth only, not for previews.
Sign-in and keys	The MCP tools use your signed-in Convex session, per SKILL.md ("cloud+user-auth only").
Scopes	Reads usage insights, table row counts and function metadata from a Convex deployment through the Convex MCP tools

Questions

What does convex-cost do?

It makes Claude read your deployment's usage evidence, rank the functions that read the most data, show how their cost grows and suggest the cheapest fix for each. It is aimed at cost surprises that come from a few read-heavy functions.

Will it tell me my bill in dollars?

No. SKILL.md asks for growth shapes and relative guidance, because pricing changes and depends on plan, and tells Claude to cite the pricing page for absolute numbers.

Does it change my code?

SKILL.md describes the work as read-only over dev and prod. It names the fix and points to convex-expert and convex-advisor for the actual change.

What does it need?

The Convex MCP tools insights, tables and functionSpec. SKILL.md says insights works on cloud deployments with a signed-in user, not on previews. The install box lists the MCP server command from the Convex docs.

How do I install it?

Use the install box at the top of this page. The first command installs convex-cost together with convex-deploy-guard, which SKILL.md uses to classify the target. convex-advisor and convex-expert are optional hand-offs for the fixes; install them separately if you want them. Run it with /convex-cost in Claude Code or $convex-cost in Codex.

Is it safe?

SKILL.md describes the cost read as read-only over dev and prod. It reads usage data and row counts through the Convex MCP tools, and the Convex docs say production access there is limited by default. Our advice: run it against dev first.

How is it different from convex-advisor?

SKILL.md says the read-heavy patterns are the ones convex-advisor flags for performance. convex-cost looks at the same evidence as money and adds the confirm-cost step for paid actions.

Is it official?

It is in the convex/agent-skills repo under the Convex GitHub organisation, which is the vendor's own repo.

Information

updated Nov 2, 2026

Repository: convex/agent-skills
Folder: skills/convex-cost

convex/agent-skills skills.sh

Sources

Docs, issue trackers and discussions this page is built from.

Show all 5

convex/agent-skills README · github.com
skills CLI commands · github.com
Convex docs: Agent Skills · docs.convex.dev
convex-cost SKILL.md · github.com
Convex docs: MCP server · docs.convex.dev