What is caveman-loop?

A step-by-step workflow for taking a software task from requirements and planning through implementation, verification, and a final report.

Use it to collect requirements, research options, plan and assess risks, make changes, verify the result, diagnose remaining problems, and repeat the cycle when needed.

It provides a repeatable process for complex work and defines completion in terms of checkable evidence instead of assumptions.

caveman-loop — Skill for Claude Code and Codex by aidvizhhub

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

Any agent

npx skills add aidvizhhub/caveman-dvizh-skills --skill caveman-loop

Clone the repo

git clone --depth 1 https://github.com/aidvizhhub/caveman-dvizh-skills

Made for: Claude Code, Codex.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

README.md

[![agentmods](https://agentmods.dev/badge/skills/aidvizhhub/caveman-dvizh-skills/caveman-loop/github.svg)](https://agentmods.dev/skills/aidvizhhub/caveman-dvizh-skills/caveman-loop)

Your own site

<a href="https://agentmods.dev/skills/aidvizhhub/caveman-dvizh-skills/caveman-loop"><img src="https://agentmods.dev/badge/skills/aidvizhhub/caveman-dvizh-skills/caveman-loop/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

Your own site · 80×15

<a href="https://agentmods.dev/skills/aidvizhhub/caveman-dvizh-skills/caveman-loop"><img src="https://agentmods.dev/badge/skills/aidvizhhub/caveman-dvizh-skills/caveman-loop.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 318 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.

When invoked 2,223 The whole file, excluding the scripts and references it only reads on demand.

Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.

Origin original No closer match found in the catalogue.

Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

Model	Per session	Once invoked
Fable 5.1	$0.00318	$0.02223
Opus 5.5	$0.00127	$0.00889
Sonnet 5.5	$0.00064	$0.00445
Haiku 4.5	$0.00032	$0.00222

Measured 2mo ago against content hash 73cc6f19726d, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

caveman-loop scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2mo ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

skills/caveman-loop/SKILL.md · 118 lines

How it starts

The opening of the file, as written. The whole thing — 118 lines — stays where its author put it; the contents beside it link to each section on GitHub.

caveman-loop (модуль полного канона)

🇬🇧 English instructions: references/EN.md — полный перевод тела.

Это часть устава пещерника. Полный канон целиком — скилл caveman-canon. Всё ниже — выдержки из BRO.md один в один (без сокращений).

⚡ ЗАКОН СКИЛЛОВ (авг 2026): на КАЖДЫЙ промт грузи как можно больше скиллов племени — МИНИМУМ 6, лучше ВСЕ 17 (canon + conclusion + teach + cheat + advise + prevent + verify + loop + kill + code + style + think + research + run + diagnose + save + desktop). Один скилл = одна грань канона; без полного груза решение кривое, охота слепая. Большая задача = ЦИКЛ ПЛЕМЕНИ (скилл caveman-loop): 8 шагов, пока «готово» не доказано фактами, кап 3 итерации → стоп и отчёт.

⚡ ДОКИ ЖЕЛЕЗНО (закон 25, авг 2026): инструмент/флаг незнакомый → СНАЧАЛА официальная дока + man + --help (RTFM-канон: первая линия, не запасная; −40% времени, −70% ошибок; официальная дока раньше чужого веба). Прогрессивно: --help → man → полная дока/вики → потом чужое; сверять с установленной ВЕРСИЕЙ. Пример: ffmpeg — man ffmpeg + ffmpeg-all.html.

СУТЬ: один цикл вместо хаоса (закон 37, авг 2026) — РЕШЕНО ✅

Канон (55 источников, кауфми 24 авг 2026: Microsoft Agent Framework «Agent Looping» — перезапуск агента, пока не выполнено условие завершения; VMAO arxiv 2603.11445 — Plan-Execute-Verify-Replan цикл; ralph-loop — один шаг → обновить спеку → закоммитить → выйти; Verces/agentnative — терминальные условия: verifier done-checks, кап итераций, no-progress стоп; promptabcd — «iteration limit is a backstop, not a strategy»; evaluator- optimizer (Vercel); opencode.ai/docs/skills — discovery/activation/execution (прогрессивное раскрытие, ~100 токенов на скилл в описании); developersdigest/Atlan/IBM/betterclaw — Skills = «как делать», MCP = «что достать»: наши скиллы — знания, НЕ тулы, MCP-обёртка не нужна (24%+ окна на определения тулов); tianpan/bestaiweb/attention-dilution — длинные промпты = распыление, поэтому ритуал КОРОТКИЙ, а детали — в скиллах):

Смысл: 13 скиллов — грани одного камня. Чтобы они работали КАК ОДНО, есть один ритуал-цикл: шаг за шагом, пока «готово» не доказано фактами. Цикл не новый сервер, а порядок + терминальный гейт + кап итераций.

Read the full file on GitHub · 118 lines

Files

What ships with it

1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

references/EN.md 5.2 KB

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

2mo ago First seen · 118 lines · 318 tokens per session scan A 73cc6f19726d

Subscribe to this mod's changes

caveman-loop is a skill published in the GitHub repository aidvizhhub/caveman-dvizh-skills (7 stars, last pushed 2mo ago), licensed MIT. It adds 318 tokens to every session and 2,223 once invoked, about $0.0013 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Claude Code Planning agent-skills ai-agents opencode skills

caveman-loop next to 6 skills from other repositories
Mod	Type	Stars	Last push	Licence	Scan	Tokens	Works with
caveman-loopaidvizhhub/caveman-dvizh-skills	Skill	7	2mo ago	MIT	A	318	Claude Code, Codex
seats-aeroborski/travel-hacking-toolkit	Skill	906	2d ago	MIT	A	78	Claude Code
biltborski/travel-hacking-toolkit	Skill	906	2d ago	MIT	A	44	Claude Code
round-the-worldborski/travel-hacking-toolkit	Skill	906	2d ago	MIT	A	51	Claude Code
scandinavia-transitborski/travel-hacking-toolkit	Skill	906	2d ago	MIT	A	49	Claude Code
southwestborski/travel-hacking-toolkit	Skill	906	2d ago	MIT	A	40	Claude Code
status-matchborski/travel-hacking-toolkit	Skill	906	2d ago	MIT	A	48	Claude Code