context-budget

To reduce Gemini CLI token consumption, audit context bloat across agents, skills, rules, MCP servers, and GEMINI.md files. This identifies redundant content and flags heavy components, producing a prioritized report of token-savings recommendations to reclaim space.

The best way to audit MCP server overhead is to estimate tool schema sizing and detect servers wrapping tools available for free. An inventory scan flags bloated frontmatter and overlapping rule content, generating actionable recommendations to reduce repeated context inflation.

Gemini CLI sessions slow down during post-expansion bloat due to excessive prompt overhead from loaded components. Analyzing token consumption across agents, skills, and MCP servers identifies duplicate skill copies and redundant rules causing the degradation.

Yes, you can detect duplicate skills and overlapping rules by performing a context overhead inventory. This heuristic token estimation flags bloated frontmatter and redundant content, allowing you to eliminate repeated context and optimize the configuration.

You need to run a context budget audit during session slowdowns, after post-expansion bloat, or pre-planning when adding new agents, skills, or MCP servers. This prevents configuration degradation by identifying heavy files before they impact performance.