Works with the AI Agents you already use
skill slim
Audit, optimize, and slim down AI agent context by identifying unused skills and calculating token costs.
Free
skill slim
Example session with this skill installed
My Claude Code agent is getting expensive and slow. Can you check if I have too many skills loaded and see what I can remove?
- Read your context and instructions
- Compiled the slim
I have performed a health check on your agent environment. You are currently loading 42 skills across Claude Code and Cursor, costing approximately 12,400 tokens in "always-on" overhead per turn.
Agent Health Audit
| Metric | Value | Status |
|---|---|---|
| Health Score | 64/100 (Grade D) | Needs Optimization |
| Always-On Tokens | 12,400 | High |
| Monthly Est. Cost | $42.50 | Based on 1k turns |
| Unused Skills | 18 | Detected in last 30 days |
| Duplicate Triggers | 4 | Conflicting descriptions |
Optimization Plan
The following skills have not been invoked in the last 30 days and are candidates for quarantine:
image-optimizer-v2(Global) - 1,200 tokenslegacy-deploy-script(Project) - 850 tokensunused-jira-helper(Global) - 2,100 tokens
Next steps
- Run
node src/cli.ts fixto see the full breakdown of recoverable tokens. - Execute
node src/cli.ts fix --applyto move these 18 skills to quarantine. - Use
node src/cli.ts serveto view the full conflict map in the web dashboard.
Connects securely to your tools. The creator never sees your data.
About this skill
AI agents often slow down or hit context limits because every installed skill and rule is loaded into every conversation turn. This bloat drives up latency and token costs without adding value. This skill audits your agent environment to identify dead weight and recover context space.
What it does
- Multi-agent auditing scans Claude Code, Codex CLI, Gemini CLI, and Cursor configurations.
- Cost calculation computes always-on token counts and estimates monthly dollar spend based on usage.
- Conflict detection identifies duplicate triggers and overlapping descriptions that confuse model routing.
- Usage tracking analyzes session transcripts to find skills that have never been invoked.
- Safe quarantine moves unused skills to a reversible storage area with a manifest for easy restoration.
- Visual reporting generates standalone HTML reports or launches a local web dashboard.
How it works
- Scan the local system using
node src/cli.tsto detect installed agents and calculate the current health score. - Analyze usage history over a set window (default 30 days) to identify never-invoked skills.
- Preview an optimization plan with
node src/cli.ts fixto see potential token and dollar savings. - Optimize the environment by applying a safe quarantine to user-owned custom skills.
Frameworks & tools
Built for Node.js (v22.6.0+) with zero external dependencies. Supports Claude Code, Codex CLI, Gemini CLI (Antigravity), and Cursor (.mdc rules).
Why this beats prompting it yourself
Manual context management is error-prone and lacks visibility into actual token costs. This tool provides hard metrics on dollar spend and uses local transcript analysis to prove a skill is unused before suggesting removal.
Use cases
- Reduce high API bills by removing always-on system instructions that are rarely used.
- Fix "tool confusion" where an agent struggles to pick between skills with overlapping triggers.
- Audit a new project's
.cursor/rulesto ensure only relevant logic is active. - Generate a health report for a complex multi-agent CLI setup.
Known limitations
Cannot move or delete plugin-bundled skills. Requires Node.js 22+ to run the TypeScript source natively. Quarantine is limited to user-owned custom skills.
How to use
Skills on Agensi run inside your AI agent. Copy the prompt, paste it into your agent, and it connects to Agensi and installs this skill for you. If a payment is needed, your agent shows it to you first.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill passes our trust checks below.
Security scanned
Passed our 8-point scan before listing
2 installs
Downloaded by developers to date
Free forever
No account required to browse
Trust & safety
Security scanned
Verified clean 3 days ago
- Free to download with an account
Requires