Skip to content

cli-reference.md

Latest commit

 

History

History
251 lines (185 loc) · 4.6 KB

cli-reference.md

File metadata and controls

251 lines (185 loc) · 4.6 KB

CLI Reference

This page is the command reference for the current Basilisk CLI surface.

Top-Level Commands

Current commands:

  • scan
  • recon
  • replay
  • modules
  • probes
  • interactive
  • sessions
  • diff
  • posture
  • version
  • help
  • eval

basilisk scan

Run a full scan against a target.

basilisk scan [OPTIONS]

Important options:

Option Description
-t, --target Target URL or API endpoint
-p, --provider Target provider
-m, --model Target model override
-k, --api-key API key file reference such as @/path/to/key
--auth Authorization header value
--mode quick, standard, deep, stealth, chaos
--evolve / --no-evolve Enable or disable the evolution engine
--generations Evolution generations
--module Repeatable module selector
--skip-recon Skip reconnaissance
--recon-module Repeatable recon module selector
--attacker-provider Provider for mutation generation
--attacker-model Model for mutation generation
--attacker-api-key Attacker-model key file reference
--exit-on-first / --no-exit-on-first Stop evolution after first breakthrough
--diversity-mode off, novelty, niche
--intent-weight Intent preservation weight
--cache / --no-cache Enable or disable response caching
--include-research-modules / --no-include-research-modules Opt into research-tier modules
--execution-mode recon, validate, exploit_chain, research
--campaign-name Campaign name
--operator Operator name
--ticket Authorization or tracking ticket
--approval-required / --no-approval-required Require explicit approval
--approve / --no-approve Mark approval confirmed
--dry-run Plan only
-o, --output html, json, sarif, markdown, pdf
--output-dir Report output directory
--no-dashboard Disable the dashboard path
--fail-on Threshold for non-zero exit
-v, --verbose Verbose output
--debug Debug mode
-c, --config YAML config file

basilisk recon

Run reconnaissance only.

basilisk recon \
  --target https://example.test/v1/chat/completions \
  --provider openai

Options:

  • --target
  • --provider
  • --api-key
  • --auth
  • --recon-module
  • --verbose

basilisk posture

Run the posture workflow.

basilisk posture \
  --provider openai \
  --model gpt-4o

Options:

  • --target
  • --provider
  • --model
  • --api-key
  • --auth
  • --output-dir
  • --json
  • --verbose

basilisk diff

Compare multiple targets side by side.

basilisk diff \
  -t openai:gpt-4o \
  -t anthropic:claude-3-5-sonnet-20241022

Options:

  • --target repeatable provider:model
  • --api-key
  • --category
  • --output-dir
  • --verbose

basilisk eval

Run an assertion-based eval suite from YAML.

basilisk eval evals/guardrails.yaml

Options:

  • --format with console, json, junit, markdown
  • --output
  • --fail-on with any or all
  • --parallel / --no-parallel
  • --diff
  • --tag
  • --verbose

basilisk modules

Inspect the attack catalog.

basilisk modules

Options:

  • --category
  • --include-research / --no-include-research
  • --json

basilisk probes

Browse the YAML probe corpus.

basilisk probes --stats

Options:

  • --category
  • --tag
  • --severity
  • --query
  • --count
  • --json
  • --stats

basilisk interactive

Start the assisted REPL:

basilisk interactive \
  --target https://example.test/v1/chat/completions \
  --provider openai

basilisk sessions

List stored sessions:

basilisk sessions

Option:

  • --db

basilisk replay

Replay one stored session:

basilisk replay BSLK-2026-ABC123

Option:

  • --db

basilisk help

Extended help topics:

basilisk help overview
basilisk help scan
basilisk help modules
basilisk help evolution
basilisk help diff
basilisk help examples

basilisk version

Show Basilisk version and local runtime information.

Environment Variables

Common provider variables:

  • OPENAI_API_KEY
  • ANTHROPIC_API_KEY
  • GOOGLE_API_KEY
  • AZURE_API_KEY
  • GH_MODELS_TOKEN
  • GROQ_API_KEY
  • XAI_API_KEY
  • BASILISK_API_KEY

Exit Behavior

For scan:

  • non-zero exit depends on --fail-on

For eval:

  • --fail-on any returns non-zero if any test fails
  • --fail-on all returns non-zero only if all tests fail

Use the advanced guides for workflow advice and the architecture pages for internals.