If you discover a security vulnerability in leash-secrets, please report it responsibly.
DO NOT open a public issue for security vulnerabilities.
Email: security@leash-secrets.dev (or open a private security advisory on GitHub)
We will:
- Acknowledge receipt within 48 hours
- Investigate and provide a timeline within 1 week
- Release a fix and credit you (unless you prefer anonymity)
Leash Secrets is a prompt-based skill and a pattern library. Its attack surface is minimal:
| Component | Risk | Mitigation |
|---|---|---|
Skill files (.md) |
Prompt injection via malicious skill content | Only install from trusted sources (this repo) |
Pattern files (.json) |
Malicious regex (ReDoS) | Patterns are reviewed; use timeout in custom implementations |
Install script (install.sh) |
Supply chain attack | Pin to specific commits; verify checksums |
| Pre-commit hook | Bypass via --no-verify |
Defense in depth; use CI scanning too |
Leash Secrets makes zero network calls after installation:
- No telemetry
- No analytics
- No crash reporting
- No update checks
- No phone-home behavior of any kind
Install-time network calls:
curltoraw.githubusercontent.comto download skill/rule filesgemini extensions installcalls GitHub (if Gemini CLI is being configured)
These are the only network calls leash-secrets ever makes, and they only happen during installation.
Leash Secrets' own skill files and patterns never contain real secrets. All examples use clearly fake values.
When leash-secrets detects a secret in your code:
- The full secret is never logged, stored, or transmitted
- Output shows only the first 6 and last 4 characters
- The pre-commit hook processes files in memory and discards findings after display
| Version | Supported |
|---|---|
| 1.x | ✅ Current |
- All releases are tagged and signed
- The install script fetches from
raw.githubusercontent.com(GitHub's CDN) - For maximum security, clone the repo and copy files manually instead of using the install script