SKILL.md
Token Saver Context Compression
Alias for context-compressor. Provides context window optimization by compressing large payloads before reasoning.
When to Use
- Context approaching budget limits (80K+ tokens)
- Large file reads that can be summarized
- Multi-agent pipelines with context handoff
- Any scenario where
context-compressorwould be used
Activation
Activate this skill when compression-reminder.txt appears or when the active context is approaching the 80K token pressure threshold.
Usage
Skill({ skill: 'context-compressor' });
This skill exists as a named alias so agents referencing token-saver-context-compression in their skills arrays resolve correctly. The implementation lives in context-compressor.