No red flags found in any of the 11 categories — no credential harvesting, no data exfiltration, no curl-pipe-shell installer. Scanned against the SlowMist agent-security taxonomy, refreshed every 8 hours. Full audit →
by Mibayy · MCP Server · ★ 1.1k
Last updated: · Indexed by AgentSkillsHub · Auto-synced every 8h
🔒 Is token-savior safe to install? View the security audit →
⚡ Token Savior Recall 97% token reduction · Persistent memory · 98 MCP tools · Python 3.11+ []() []() []() []() What it does Token Savior Recall is a Claude Code MCP server that solves two problems: 1. Token waste — Claude reads entire files to answer questions about 3 lines. Token Savior navigates your codebase by symbols, returning only what's needed. 97% reduction on 170+ real sessions. 2. Amnesia — Claude starts from zero every session. Token Savior Recall captures observations across sessions, injects relevant context at startup, and surfaces the right knowledge before you ask. findsymbol("sendmessage")
| Stars | 1,110 |
| Forks | 94 |
| Language | Python |
| Category | MCP Server |
| License | MIT |
| Quality Score | 67.6437954314905/100 |
| Last Updated | 2026-08-10 |
| Created | 2026-03-30 |
| Platforms | claude-code, cli, mcp, python |
| Est. Tokens | ~19k |
Explore other popular mcp server tools:
token-savior is MCP server that gets Claude to 97.9% (188/192) on a real coding benchmark at -80% active tokens and -83% wall time, vs 78.3% plain. Structural code navigation + persistent memory engine. Works with ev. It is categorized as a MCP Server with 1.1k GitHub stars.
token-savior is primarily written in Python.
You can find installation instructions and usage details in the token-savior GitHub repository at github.com/Mibayy/token-savior. The project has 1.1k stars and 94 forks, indicating an active community.
token-savior is released under the MIT license, making it free to use and modify according to the license terms.
Grades come from a rule-based scan built on the SlowMist agent-security taxonomy, covering 11 red-flag categories including credential harvesting, data exfiltration, and curl | sh installers. It is a first-layer scan, not a manual audit — we say so rather than overstate it.
The scale of the problem is documented independently: Liu et al. (2026), in a study of 31,132 agent skills, report that 26.1% contain security vulnerabilities. Our own full-catalog census is published as a citable open dataset.
Sources & who's responsible: