What problem does it solve?
This Skill addresses the challenge of excessive token usage in AI agent sessions, particularly those exceeding context limits or involving large codebases. It aims to optimize token usage without sacrificing critical information.
Core Features & Use Cases
- Context Compression Strategies: Offers three approaches for compression - anchored iterative summarization, opaque compression, and regenerative full summary - tailored to different session characteristics.
- Tokens-Per-Task Optimization: Focuses on the total tokens consumed to complete a task, rather than individual requests, to ensure efficiency.
- Artifact Trail Integrity: Ensures that key details such as file modifications and error messages are preserved for debugging and artifact tracking.
- Structured Summaries: Generates detailed, structured summaries with mandatory sections to prevent silent information loss.
Quick Start
Run the context-compression skill for the current session to apply the optimal compression strategy based on the session characteristics.