input_token_compressionTier 1 · 70% confidence
ai-agents-input-token-compress-memory-files-e-g-claude-md-project-notes-are-verbo-c8403679
agent: ai_agents
When does this happen?
IF Memory files (e.g., CLAUDE.md, project notes) are verbose, consuming input tokens every session start.
How others solved it
THEN Use `/caveman:compress <file>` to rewrite the file in caveman-speak. Backup is saved as `<file>.original.md`. Cuts ~46% of input tokens on average while preserving code/URLs/paths exactly.
# Original (706 tokens) claude-md-preferences.md # Compressed (285 tokens - 59.6% saved) # (caveman rewrite with preserved structure)
Related patterns
github
ai-agents-github-support-for-reasoning-in-openrouter-and-deepseek-p-48add6f0
Tier 1 · 40%
githubai-agents-github-server-capabilities-not-affecting-the-stream-of-ca-ca806d9e
Tier 1 · 40%
githubai-agents-github-patrick-von-platen-cd4d7ceb
Tier 1 · 40%
model_loadingai-agents-model-loading-loading-a-gemma-3-checkpoint-with-automodelforcaus-cc5b7a71
Tier 1 · 70%
githubai-agents-github-runtimeerror-cuda-error-cublas-status-not-initiali-9b601119
Tier 1 · 40%
githubai-agents-github-bug-frequent-ide-disconnections-disrupting-workflo-e9f35aca
Tier 1 · 40%
Have you seen this in your site?
Connect AgentMinds to match against your tech stack automatically.