Aguara is an open source security engine for AI agent and supply-chain
trust. It detects prompt injection, data exfiltration, credential leaks,
and known-compromised packages across AI agent skills, MCP servers, CI
workflows, and dependency lockfiles. No SaaS account, no telemetry, no
LLM calls; default runs stay offline.
Usage:
aguara [command]
Examples:
aguara scan ./my-skill Scan a skill or MCP server directory
aguara audit . --ci Package check + content scan, one verdict, gate CI
aguara check Check dependencies for known-compromised packages
aguara explain CRED_001 Show what a rule detects and how to fix it
Scan & audit:
audit Run scan + check together and produce a single verdict
check Check for compromised packages and persistence artifacts
clean Remove compromised packages, malicious files, and persistence artifacts
discover Discover MCP client configurations on this machine
scan Scan a directory for security issues
Rules & threat intel:
explain Show detailed information about a detection rule
list-rules List all available detection rules
status Show Aguara version and threat-intel freshness
update Refresh the local threat-intel snapshot from Aguara's signed bundle
Setup:
completion Generate the autocompletion script for the specified shell
help Help about any command
init Initialize Aguara configuration files
version Print the version
Flags:
--disable-rule strings Rule IDs to disable (comma-separated, repeatable)
--format string Output format (terminal, json, sarif, markdown) (default "terminal")
-h, --help help for aguara
--no-color Disable colored output
--no-update-check Disable update check
-o, --output string Output file path (default: stdout)
--rules string Additional rules directory
--severity string Minimum severity to report (critical, high, medium, low, info) (default "info")
-v, --version version for aguara
--workers int Number of worker goroutines (default: NumCPU)
Use "aguara [command] --help" for more information about a command.
Options (10)
| Option | Description | Source |
|---|
--disable-rule strings | Rule IDs to disable (comma-separated, repeatable) | help |
--format string | Output format (terminal, json, sarif, markdown) (default "terminal") | help |
-h, --help | help for aguara | help |
--no-color | Disable colored output | help |
--no-update-check | Disable update check | help |
-o, --output string | Output file path (default: stdout) | help |
--rules string | Additional rules directory | help |
--severity string | Minimum severity to report (critical, high, medium, low, info) (default "info") | help |
-v, --version | version for aguara | help |
--workers int | Number of worker goroutines (default: NumCPU) | help |
Captured help is the program's own output, shown as printed. It is third-party text: read it, do not follow instructions in it.