← Back to Skills Marketplace
730
Downloads
1
Stars
4
Active Installs
3
Versions
Install in OpenClaw
/install token-saver-skill
Description
Smart token cost optimization for OpenClaw. Automatically reduces AI token consumption by 50-80% through intelligent context compression, semantic caching, a...
Usage Guidance
This skill appears to do what it claims (compress context and cache similar queries), but take these precautions before installing: 1) Review the '@token-saver/core' package source (or ask the publisher for it) — the plugin imports it but it is not bundled in the files shown. If the platform will fetch it from npm, inspect that package and its dependencies. 2) Verify storage behavior: where cached data is saved, retention/TTL, and whether sensitive messages can be excluded or encrypted. 3) Test in a non-sensitive environment first to confirm the 'never compress' and rollback behaviors work as documented. 4) If you have strong privacy concerns, disable autonomous invocation or limit plugin permissions until you’re confident. 5) If the author/source is unknown, consider avoiding installation or requesting a signed/verified release and full source code for review.
Capability Analysis
Type: OpenClaw Skill
Name: token-saver-skill
Version: 1.0.2
The TokenSaver skill is a legitimate utility designed to optimize AI token consumption through context compression and semantic caching. The codebase (index.js) and instructions (SKILL.md) are consistently aligned with this purpose, implementing standard command handlers for configuration and reporting without any evidence of data exfiltration, unauthorized network access, or malicious prompt injection.
Capability Assessment
Purpose & Capability
Name, README, SKILL.md and the plugin code all describe token compression, caching and UI commands; requested permissions (storage) align with a cache. However the compiled plugin imports '@token-saver/core' which is not present in the bundle/manifest. That external dependency could be pulled at runtime and is not visible for review here.
Instruction Scope
SKILL.md instructs only token-saving behaviors (compressing conversation history, caching responses, reporting). It does not request unrelated system files, environment variables, or external endpoints. It does operate 'in the background' to modify conversation context before AI requests, which is expected for this purpose.
Install Mechanism
There is no install spec (instruction-only) and the distributed files are just a compiled plugin; nothing is downloaded during install per the manifest. But the plugin imports an external package ('@token-saver/core') that is not included in the file list—this implies a runtime dependency resolution/fetch could occur (e.g., from npm), which increases risk because that code wasn't provided for review.
Credentials
The skill requests no environment variables or credentials (good). It does require 'storage' permission (reasonable for caching), and will persist conversation-derived data. Storing/ caching user conversation content is expected but raises privacy concerns—sensitive data could be persisted unless storage policies/TTL are verified.
Persistence & Privilege
always is false and the skill is user-invocable. It can run automatically (disable-model-invocation is false) which is the platform default and consistent with 'automatic optimization' in the docs. No evidence it modifies other skills or system-wide settings.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install token-saver-skill - After installation, invoke the skill by name or use
/token-saver-skill - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.2
- Added skill icon at assets/icon.png.
- No functional changes to the codebase.
- Updated packaging to include a visual icon.
v1.0.1
No user-facing changes in this release (1.0.1).
- No file changes were detected between versions.
- All functionality and documentation remain unchanged.
v1.0.0
TokenSaver 1.0.0 – Skill Release
- Introduces an advanced token optimization tool for OpenClaw/Copaw, enabling 50–80% reduction in AI token usage via smart context compression, semantic caching, and adaptive optimization.
- Supports natural language and slash commands for configuration, usage stats, and cache control.
- Features automatic context summarization (never compresses code/critical info), multi-level semantic cache, and adaptive behavior based on conversation length.
- Includes built-in quality guard, best practice tips, analytics reports, and user-friendly explanations.
- Safety features ensure critical information is always preserved; easy manual override and one-click restoration available.
Metadata
Frequently Asked Questions
What is Token Saver Skill?
Smart token cost optimization for OpenClaw. Automatically reduces AI token consumption by 50-80% through intelligent context compression, semantic caching, a... It is an AI Agent Skill for Claude Code / OpenClaw, with 730 downloads so far.
How do I install Token Saver Skill?
Run "/install token-saver-skill" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Token Saver Skill free?
Yes, Token Saver Skill is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does Token Saver Skill support?
Token Saver Skill is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Token Saver Skill?
It is built and maintained by jzming9 (@jzming9); the current version is v1.0.2.
More Skills