@juliusbrussee

caveman-compress reduces token usage by collapsing natural language files into terse, semantically equivalent form while preserving code, urls, and technical content exactly. runs via cli with automated backup and retry logic.

read-only drift detector that compares spec against code, classifies violations by type (violate, drift, missing, stale, unverifiable), and reports grouped by severity with remediation hints. writes nothing.

caveman-review strips filler from code review feedback, delivering compressed single-line comments with location, problem, and fix. targets noise reduction in pull request workflows.

cavecrew routes code tasks to three compressed subagents (investigator, builder, reviewer) that reduce context cost per delegation by ~60%, trading prose for structured fragments in long sessions where token budget is tight.

caveman compresses responses by dropping articles, filler, and pleasantries while preserving technical accuracy and code. toggleable via command with intensity levels. designed for token reduction in longer conversations.

spec is an incomplete skill mutator for managing SPEC.md files with four modes (new, distill, backprop, amend) but lacks clear procedural definition, trigger examples, and error handling.

caveman-es is a style guide for terse spanish technical communication, not an executable skill. it prescribes tone/format rules without defining any actionable procedure, inputs, or outcomes.

caveman is a style guide for technical writing that prescribes grammar rules, symbol preferences, and structural templates. it is not an executable skill for ai agents.

Caveman in Classical Chinese: 文言文 register, far fewer characters, technical terms verbatim. Invoke only with /megacave or /caveman wenyan. Stays on until "stop…

Caveman at maximum compression: fragments, one word when one word is enough, each fact once. Invoke only with /ultracave or /caveman ultra. Stays on until…

Compress natural language memory files (CLAUDE.md, todos, preferences) into caveman format to save input tokens. Preserves all technical substance, code, URLs,…

Wire a repository through the Caveman Cloud gateway so every LLM request is measured, with no behavior change. Use for "set up caveman" or adding LLM spend…

Implement reversible compatibility-safe transitions. Use for schema, data, API, protocol, configuration, or dependency migrations requiring rollback and…

Inspect Caveman Cloud's experiment lifecycle and block unsafe execution. Use when asked to start, approve, cancel, promote or roll back a Caveman experiment.

Find and label every LLM workflow in the repository so Caveman Cloud groups spend by workflow instead of one bucket. Use for "discover workflows" or breaking…

Build feature work with high overbuilding risk. Use for new behavior, product slices, or integrations where repository reuse, strict scope, and an explicit…

Read-only review of Caveman Cloud evidence: cost, Cave Score, workflows, traces, latency, errors, routing, savings. Use when asked what Caveman found or where…

Fix bugs and small behavior changes at the narrowest responsible layer. Use when regression proof, preserved surrounding behavior, and task-relevant tests…

Prove existing work meets acceptance conditions without expanding scope. Use for validation-only tasks, completion checks, focused gate runs, and last-mile…

Diagnose ambiguous failures before editing. Use for unknown causes, intermittent behavior, performance regressions, or investigations needing evidence-ranked…

Turn a Caveman optimization observation into an operator-chosen candidate with a paired baseline evaluation. Use when asked to inspect or evaluate a Caveman…

Act on a Caveman learn report - review the ranked token sinks, apply cost-lowering fixes with per-edit consent, and report what those fixes returned. Use when…

Restructure code while preserving behavior. Use for extraction, consolidation, ownership moves, or cleanup where verification must bracket structural edits.

Read-only repository explorer for cold-start orientation, broad cross-file localization, or when a direct search failed. Skip it when the exact file or symbol…

Show recorded output and cache-read token usage and mode attribution for the current Claude Code session, or locate the host's native usage report. Trigger:…

Quick-reference card for the three caveman skills and their commands. Trigger: /caveman-help or "caveman help".

Write a Conventional Commits message compressed to intent only. Use for "write a commit", "commit message", /commit or /caveman-commit.

Terse caveman voice: answer first, fluff gone, every technical fact kept. Use for /caveman, "caveman mode", "talk like caveman", "be brief", "less tokens".…

Plan-then-execute implementation against SPEC.md. Native single-thread loop, no sub-agents. On test or build failure, auto-invokes the backprop skill before…

Bug → spec protocol. When a bug is found or a test fails, trace the cause, decide whether a new §V invariant would catch recurrence, append to §B. This is the…

>