TESIGN
SAVED

EN · KO

  1. TESIGN / RADAR
  2. caveman

caveman

Token savings for agent users

WHO USES ITSolo developer · Team manager

For developers

Cuts an agent's token use by stripping prompts down to bare words.

★ 106k▲ +222 / 7dDaily star increases over 14 days. Hollow bars mean no data. The last day may be partial.2026-09-03 +62026-09-04 +1402026-09-05 +972026-09-06 +652026-09-07 +62026-09-08 +22026-09-09 +22026-09-10 +102026-09-11 +862026-09-12 +582026-09-13 +222026-09-14 +42026-09-15 +192026-09-16 +23largest one-day increase in the 14 days+140last 14 days · stars per day

105,788 stars · checked on GitHub GitHub check 30 h late · incl. +28 observed via GH Archive since then · 7-day +222 observed via GH Archive (as of )

RANKS All-time #32 Rising this week #49 DEV TOOLS #6

The 'Caveman Learn' report screen: saved-token figures and bars, a list of token sinks and a session-context-depth bar chart.
The 'Caveman Learn' report screen: saved-token figures and bars, a list of token sinks and a session-context-depth bar chart.

TESIGN TAKE

A skill for output and a proxy for input; admitting that savings vary by case is what makes it credible.

USE PACKAGE

AT A GLANCE

USAGE
Unconfirmed — check before use.
LANGUAGE
Go
PLATFORM
cli
ACTIVITY
last commit 1 day ago () · latest release v2.7.0 () · 35 releases
COMMUNITY
56 contributors · 119 open issues (incl. PRs) · made by: an individual
SOURCES
GitHub
FIRST SEEN
CATEGORY
AI · DEV TOOLS

A summary, not legal advice.

WHY IT MATTERS

The longer an agent's replies, the more tokens and cost go with them. caveman is a Claude Code skill that strips the agent's replies down to bare words, so the same content comes back in far less text. The repository claims a 65% token reduction, but that figure is its own. Written in Go; the skill is MIT and the bundled proxy engine BSL-1.1.

BUILD FROM THIS

  • A setting that trims output tokens on repeat jobs with long replies
  • A cost experiment that measures the saving on the team's own tasks

WHO IT'S FOR

Solo developer
sees tokens leaking into long agent replies
Team manager
measures, then trims, token cost on repeat jobs

START IN 5 MINUTES

# Skill (MIT):
$ npx skills add JuliusBrussee/caveman
# Proxy (Node.js 22.13+):
$ npm install -g @caveman-ai/cli && caveman setup --install
$ caveman claude

CAVEATS

  • The skill only shortens output, and its rules cost 1 to 1.5k input tokens per turn
  • On already-terse work it can cost more, as the README says
  • The skill is MIT; the proxy engine is BSL-1.1 (hosting for others needs a licence)

RECEIPT

FIRST SEEN
AT SOURCE
KEPT
STAR DELTA
+23 ★ / 24h
CREATED → FIRST SEEN
161d
WHEN FIRST SEEN
105,284 ★
SOURCES
GitHub

The same facts in machine-readable form — View as Markdown · JSON

SIMILAR TOOLS

Up to five from AI by ★ total: edited entries first, then repository cards; this entry is highlighted. Drawn from the same stored snapshot as the rankings — a comparison, not a recommendation.

IMAGENAME★ TOTAL▲ 7dLICENSEPLATFORMLAST PUSH
ECC README banner: headline 'The operating system for AI agent harnesses', harness chips, and skill, agent and command lists.
eccSkills, memory and security for agents★ 259k▲ +832MITpermissivewindows · macos · linux · cli
Hermes Agent landing: blue background, the headline 'The agent that grows with you' and an engraving-style winged figure.
Hermes AgentSelf-improving personal agent★ 246k▲ +375MITpermissivelinux · macos · windows · android · cli
DeepSeek Harness landing: blue background, the headline 'Everything is a plugin' and a quick-start box with the npx command.
DeepSeek HarnessPlugin agent harness for developers★ 225k▲ +1,356MITpermissiveweb · cli
The 'Caveman Learn' report screen: saved-token figures and bars, a list of token sinks and a session-context-depth bar chart.
cavemanToken savings for agent usersTHIS ENTRY★ 106k▲ +222Licence unconfirmedunconfirmedcli
mattpocock/skillsCARD★ 261k▲ +872MITpermissive—

ALSO WORTH SEEING

PodroidLinux container VM for your phonePodroid's homepage: a dark screen reading "Linux containers on Android. No root required," with a Download APK button.Install to use
★ 2.6k▲ +14 / 7d
Boots a real, separately-kernelled Alpine Linux VM on an unrooted Android phone, running Podman, Docker and LXC exactly as they behave on a server.OpenHunterAILocal AI red-team workspaceOpenHunterAI's homepage, labelled "AUTHORIZED EXTERNAL SECURITY TESTING," listing its five-stage process from scope verification to reporting.Self-host
★ 165▲ +15 / 7d
An alpha-stage local workspace that runs attacker-style security assessments — strictly against public targets you own or are authorized to test.MartinLoopExecution-control CLI for coding agentsMartinLoop's homepage headline: "Watch your AI burn $5.20. Then watch MartinLoop close the loop for $2.30."
★ 130▲ +21 / 7d
Wraps a coding agent with a spend cap and a verifier, so "I'm done" comes with an evidence-backed receipt instead of just a claim.Spec KitSpec-first process for coding agentsSpec Kit docs landing page — the headline 'Build with a spec, fix a bug, or assess an idea' and the uv tool install specify-cli commandInstall to use
★ 137k▲ +560 / 7d
Makes your coding agent write the spec first, then plan, build and verify against it.

TIMELINE

  1. Repository created
  2. FIRST SEEN BY TESIGN