Caveman Setup skill: what it does and how to install it
Wires LLM clients through the Caveman Cloud gateway to measure requests and spend without changing model behaviour.
Summary generated from the skill's documentation.
Install
$ npx skills add JuliusBrussee/caveman --skill caveman-setupRun it in a terminal. If your agent is already running, start a new session so it picks the skill up.
About this skill
What it does. Finds live LLM callsites, routes them through Caveman with the appropriate gateway and provider-key headers, and stores credentials in environment variables. It derives an app slug, supports stored or BYOK provider keys, then verifies the wiring with one small real request and reports its usage.
When to use it. Applies when adding Caveman spend observability to a repository with existing LLM calls. It stops without changes when no callsites are found and reports the integration as unverified if the gateway or provider request fails.
History
Repo stars
109kAbout +8k since 9 Jul 2026
Before 1 Oct 2026 the curve is estimated from public event data.
Stars are counted for the whole repository, which holds 20 skills.
Show as a table
| Date | Repo stars |
|---|---|
| 1 Oct 2026 | 108,672 |
| 24 Sept 2026 (estimated) | 108,359 |
| 17 Sept 2026 (estimated) | 107,939 |
| 10 Sept 2026 (estimated) | 105,130 |
| 3 Sept 2026 (estimated) | 101,650 |
| 27 Aug 2026 (estimated) | 101,470 |
| 20 Aug 2026 (estimated) | 101,320 |
| 13 Aug 2026 (estimated) | 101,310 |
| 6 Aug 2026 (estimated) | 101,120 |
| 30 Jul 2026 (estimated) | 101,070 |
| 23 Jul 2026 (estimated) | 100,990 |
| 16 Jul 2026 (estimated) | 100,920 |
| 9 Jul 2026 (estimated) | 100,711 |
Installs
76.7k
Tracking since . A chart appears once there are 7 days of data.
Installs via skills.sh
Similar skills
- Code SimplificationRefactors working code so it is easier to read, maintain and extend, without changing its behaviour.
- Requesting Code ReviewRequests a code review to verify that completed work meets its requirements before merging.
- Systematic DebuggingMakes the agent find the root cause of a bug, test failure or unexpected behaviour before it proposes a fix.