maiconfz--context-token-reduction--token-reduction-advisor

What problem does it solve? AI agent configurations often accumulate excessive always-on context that inflates token usage. This Skill turns a footprint report into an ordered, plan-only reduction strategy without touching any files, ensuring changes are reviewed before implementation. ## Core Features & Use Cases - Consent-Gated Planning: Refuses to draft a plan until the user explicitly consents, placing blocking questions when consent is missing. - Ordered Reduction Actions: Produces a markdown plan covering deletions, splits, glob-scoping, always-on to on-demand conversion, ignore rules, and IDE mirror generation, prioritized by severity. - Safety Preservation: Explicitly keeps safety, security, and secret-handling rules intact and never recommends dropping them to save tokens. - Use Case: After running a token footprint analysis on your repository's agent configuration, use this Skill to receive a prioritized, risk-aware reduction plan you can review and approve before any edits occur. ## Quick Start Ask the agent to draft a context-token reduction plan from the footprint report after confirming you consent to planning.