CLI Reference¶

Complete command-line interface reference for SuperClaw.

Global Options¶

superclaw [OPTIONS] COMMAND [ARGS]

Option	Description
`--version`	Show version and exit
`--help`	Show help message

Commands¶

attack¶

Run security attacks against an agent.

superclaw attack <adapter> [OPTIONS]

Arguments:

Argument	Description
`adapter`	Target adapter: `openclaw`, `mock`, `acp`

Options:

Option	Description	Default
`--target`	Target URL	Config default
`--behaviors`	Comma-separated behaviors to test	`all`
`--techniques`	Comma-separated attack techniques	`all`
`--output`	Save results to file	None
`--dry-run`	Preview without executing	`false`
`--timeout`	Attack timeout in seconds	`30`

Examples:

# Basic attack
superclaw attack openclaw --target ws://127.0.0.1:18789

# Specific behaviors
superclaw attack openclaw --behaviors prompt-injection-resistance,sandbox-isolation

# Save results
superclaw attack openclaw --output results.json

# Dry run
superclaw attack openclaw --dry-run

generate¶

Generate attack scenarios using Bloom.

superclaw generate scenarios [OPTIONS]

Options:

Option	Description	Default
`--behavior`	Behavior to generate scenarios for	Required
`--num-scenarios`	Number of scenarios to generate	`10`
`--variations`	Comma-separated variations	None
`--output`	Save scenarios to file	None

Examples:

# Generate prompt injection scenarios
superclaw generate scenarios --behavior prompt_injection --num-scenarios 20

# With variations
superclaw generate scenarios --behavior jailbreak --variations noise,emotional_pressure

# Save to file
superclaw generate scenarios --behavior prompt_injection --output scenarios.json

evaluate¶

Evaluate an agent against pre-generated scenarios.

superclaw evaluate <adapter> [OPTIONS]

Options:

Option	Description	Default
`--scenarios`	Path to scenarios file	Required
`--behaviors`	Behaviors to evaluate	`all`
`--output`	Save results to file	None

Examples:

superclaw evaluate openclaw --scenarios scenarios.json --behaviors all
superclaw evaluate mock --scenarios scenarios.json

audit¶

Run a comprehensive security audit.

superclaw audit <adapter> [OPTIONS]

Options:

Option	Description	Default
`--target`	Target URL	Config default
`--quick`	Quick audit (critical behaviors only)	`false`
`--comprehensive`	Full audit (all behaviors)	`false`
`--report-format`	Output format: `html`, `json`, `sarif`	`json`
`--output`	Output file path	`audit-report`

Examples:

# Quick audit
superclaw audit openclaw --quick

# Comprehensive with HTML report
superclaw audit openclaw --comprehensive --report-format html --output report

report¶

Generate and analyze reports.

superclaw report <subcommand> [OPTIONS]

Subcommands:

generate¶

Generate a report from results.

superclaw report generate --results <file> --format <format>

Option	Description
`--results`	Path to results JSON file
`--format`	Output format: `html`, `json`, `sarif`
`--output`	Output file path

drift¶

Compare two runs for regressions.

superclaw report drift --baseline <file> --current <file>

Option	Description
`--baseline`	Path to baseline results
`--current`	Path to current results
`--output`	Output file path

Examples:

# Generate SARIF for GitHub
superclaw report generate --results results.json --format sarif --output results.sarif

# Drift detection
superclaw report drift --baseline baseline.json --current current.json

scan¶

Scan for security issues.

superclaw scan <subcommand> [OPTIONS]

Subcommands:

config¶

Scan SuperClaw configuration.

superclaw scan config [--output <file>]

skills¶

Scan agent skills/plugins for risky patterns.

superclaw scan skills --path <directory>

Examples:

superclaw scan config
superclaw scan skills --path /path/to/skills

codeoptix¶

CodeOptiX integration commands.

superclaw codeoptix <subcommand> [OPTIONS]

Subcommands:

status¶

Check CodeOptiX integration status.

superclaw codeoptix status

register¶

Register SuperClaw behaviors with CodeOptiX.

superclaw codeoptix register

evaluate¶

Run multi-modal security evaluation.

superclaw codeoptix evaluate [OPTIONS]

Option	Description	Default
`--target`	Target URL	Config default
`--llm-provider`	LLM provider	`openai`
`--behaviors`	Behaviors to evaluate	`all`

Examples:

superclaw codeoptix status
superclaw codeoptix register
superclaw codeoptix evaluate --target ws://127.0.0.1:18789 --llm-provider anthropic

Utility Commands¶

behaviors¶

List all available security behaviors.

superclaw behaviors

attacks¶

List all attack techniques.

superclaw attacks

init¶

Initialize configuration file.

superclaw init

Creates ~/.superclaw/config.yaml with defaults.

Exit Codes¶

Code	Meaning
`0`	Success / All behaviors passed
`1`	Critical or high severity findings detected
`2`	Connection or runtime error

Environment Variables¶

Variable	Description
`SUPERCLAW_TARGET`	Default target URL
`SUPERCLAW_AUTH_TOKEN`	Authorization token for remote targets
`SUPERCLAW_LLM_PROVIDER`	LLM provider for Bloom
`SUPERCLAW_LOG_LEVEL`	Logging verbosity
`OPENAI_API_KEY`	OpenAI API key
`ANTHROPIC_API_KEY`	Anthropic API key