Skip to content
Get packs

Caveman Cloud Workflow Discovery

Finds LLM jobs in a repo and labels them so Caveman Cloud spend is grouped by workflow.

Source Download skill
Repository
JuliusBrussee/caveman
Stars
116,835
Licence
Apache-2.0
Files
1 file

What it does

This skill tells an AI coding agent to scan a codebase for places that call a language model, such as request handlers, scheduled jobs, CLI scripts and eval harnesses, and to give each distinct job a readable label like support-reply or nightly-digest. The labels are added as request headers or options so the Caveman Cloud gateway can show token spend per job instead of one unlabeled bucket. The agent must first show a proposed table of labels and wait for the user's approval before editing any code, then verify one labeled request still succeeds.

Great for

  • Engineering teams already routing LLM traffic through the Caveman Cloud gateway who want per-job cost breakdowns
  • Developers maintaining services with several distinct LLM jobs sharing one codebase
  • Operators who want a reviewed, minimal labeling diff rather than automatic edits

Skip it if

  • Repositories not routed through the Caveman gateway: labels only travel on gateway traffic, so the skill lists those callsites as not wired and changes nothing for them
  • Users wanting a general code audit or cost tool unrelated to Caveman Cloud, since the output is tied to its header format and dashboard
  • Teams who cannot review code diffs: the skill edits source files once approved, so careless approval could introduce bugs

Try asking

  • “Discover workflows in this repo and propose labels for LLM spend breakdown”
  • “Break down our LLM spend by workflow for the Caveman dashboard”
  • “Label the nightly digest and support reply jobs for Caveman Cloud”
Step-by-step guide: how to use Caveman Cloud Workflow DiscoveryInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 1 original files, the Apache-2.0 licence and an attribution note.

    caveman-discover.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip caveman-discover.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip caveman-discover.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Review the proposed labeling table carefully before approving; slugs are long-lived and renaming splits spend history
  • Expect scheduled jobs to appear on the dashboard only when their schedule next fires
  • Entries marked review had purposes inferred from filenames, so confirm them before relying on the report

What's inside

  • SKILL.md

Redistributed unmodified from JuliusBrussee/caveman under the Apache-2.0 licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.