Reduce coding-agent context load with Token Savior
Connect coding agents to a Token Savior MCP server for structural code navigation, persistent recall, compact command output, and token-aware project context reuse.
npx skills add agentskillexchange/skills --skill reduce-coding-agent-context-load-with-token-savior
Use Token Savior when a coding agent repeatedly reloads the same files, loses project context, or burns tokens on verbose shell and test output. The operator installs the token-savior-recall MCP package, adds it to the agent’s MCP configuration, and lets the agent use structural navigation, persistent recall, capture retrieval, and usage stats during coding sessions. Keep the scope to token-aware coding-agent navigation and recall; it is not a generic memory product, benchmark card, or replacement for normal code review and testing.
What this skill actually does
Inputs and prerequisites: MCP-compatible coding agent, Python, pip or uv, token-savior-recall package, access to local project files and optional shell/test output hooks.
Setup notes: Install with pip install “token-savior-recall[mcp]”, add the token-savior-recall server to the MCP client configuration, and optionally run ts init –agent claude –yes or the matching agent setup command documented upstream.
Source and verification boundary: use https://mibayy.github.io/token-savior/ as the canonical reference before running the workflow; keep commands, API calls, CLI usage, and generated outputs reviewable against that upstream source.
Framework fit: publish this as a MCP workflow only when the operator can invoke the documented toolchain directly, rather than treating the upstream project as a generic product listing.