More screenshots
Works with the AI Agents you already use
hetzer
Enterprise-grade secret mediation and stream redaction for AI agents to prevent credential leaks.
$9.90
hetzer
Example session with this skill installed
I need to push the current changes to the main branch on GitHub, but I don't want to expose my GITHUB_TOKEN in the logs. How do I do this safely?
- Read your context and instructions
- Compiled the hetzer
I will execute the git push using Hetzer's scoped execution to ensure your GITHUB_TOKEN remains masked and isolated from the session logs.
Secure Git Push Execution
| Step | Action | Status |
|---|---|---|
| 1 | Verify github-token in vault | hetzer_vault_has(github-token) -> true |
| 2 | Prepare scoped environment | --strict isolation enabled |
| 3 | Execute mediated command | hetzer exec --allow github-token -- git push |
| 4 | Stream Redaction | Active (sliding-window) |
# Executing scoped command...
hetzer exec --allow github-token --strict -- git push origin main
# [Hetzer] Mediation active for: github-token
# [Hetzer] Redacting stdout/stderr streams...
# Enumerating objects: 5, done.
# Counting objects: 100% (5/5), done.
# Delta compression using up to 8 threads
# Compressing objects: 100% (3/3), done.
# Writing objects: 100% (3/3), 342 bytes | 342.00 KiB/s, done.
# Total 3 (delta 2), reused 0 (delta 0), pack-reused 0
# To github.com:user/repo.git
# a1b2c3d..e4f5g6h main -> main
Next steps
- Run
hetzer_vault_list()to see other available credential scopes for this project. - If you need to add new deployment keys, run
hetzer creds set <id>in your local terminal. - Review
.hetzer/brokers/if you want to transition this to a loopback proxy for added memory safety.
Connects securely to your tools. The creator never sees your data.
About this skill
Agents handling cleartext API keys, database URLs, and tokens in shell logs or prompt histories represent a critical security vulnerability. Hetzer solves this by providing enterprise-grade secret mediation and stream redaction, ensuring models never touch raw credentials.
What it does
- Secret Masking: Replaces raw strings with
secretRefsyntax to prevent credentials from leaking into model context or logs. - Scoped Execution: Runs shell commands through a mediator that injects secrets only at the subprocess boundary.
- Stream Redaction: Continuously filters child process stdout and stderr through a sliding-window redactor to catch accidental leaks.
- HTTP Brokering: Uses loopback proxies to handle API authentication, keeping upstream keys entirely out of process memory.
- Canary Tripwires: Injects honeytokens to detect and terminate malicious scripts attempting to exfiltrate data.
How it works
- Registration: The user registers sensitive keys out-of-band using
hetzer creds setto keep them in an AES-256-GCM vault. - Reference: The agent uses the
secretRef:idsyntax in configuration files or environment variables. - Execution: The agent wraps shell commands in
hetzer exec --allow <id>to grant temporary, scoped access. - Verification: The agent uses MCP tools like
hetzer_vault_hasto check credential availability without viewing the secret itself.
Frameworks & tools
Built for Node.js environments (v22.5.0+). Compatible with MCP-enabled agents including Claude Code, Cursor, Windsurf, and Roo Code.
Why this beats prompting it yourself
System prompts cannot reliably prevent a model from echoing a variable it just read or logging it during a failed build. Hetzer enforces safety at the process level, providing a physical barrier that works even if the model ignores instructions.
Use cases
- Secure Publishing: Running
npm publishorgit pushwithout exposing tokens in the terminal scrollback. - CI/CD Automation: Mediating database migrations where the agent requires access to a production connection string.
- Untrusted Scripts: Executing third-party dependencies or external scripts within a sandboxed container environment.
Known limitations
Requires Node.js 22.5.0 or higher. It does not intercept low-level kernel APIs or memory scrapers outside the mediated process tree.
Requires Node.js 22.5.0+. Does not intercept low-level kernel APIs or memory scrapers. Security relies on same-user trust for vault key access.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 1 month ago
Requires