Agent Mode

mysecutec agent is the one-command way into AI-agent-driven use of the CLI. It:

1. Detects an installed agent — Claude Code (claude), OpenAI Codex (codex), Gemini CLI (gemini), or Cursor (cursor-agent).
2. Creates a working directory (default ./mysecutec-workdir).
3. Writes the AGENTS.md guide into it, so the agent knows how to use the CLI safely (conventions, the full command catalog, confirmation rules).
4. Launches the agent there with an initial prompt, dropping you straight into a session where you talk to your security data in natural language.

mysecutec agent

If several agents are installed, it asks which to use. If none are found it tells you what to install (or use --no-launch to just prepare the directory).

Flags

Flag	Default	Description
--agent	auto-detect	Agent to use (claude, codex, gemini, cursor-agent)
--dir	mysecutec-workdir	Working directory to create and launch in
--prompt	built-in	Initial prompt for the agent
--no-launch	false	Prepare the working directory but don't launch
--force	false	Overwrite an existing AGENTS.md in the directory
-y, --yes	false	Skip the confirmation prompt

Examples

# Auto-detect an agent and start, using the active environment
mysecutec agent

# Force a specific agent and a custom workdir + prompt
mysecutec agent --agent claude --dir ./sec-review \
  --prompt "Find my most urgent threats and summarize the dark-web exposure"

# Just set up the folder + AGENTS.md, launch the agent yourself later
mysecutec agent --no-launch

Notes

• The agent runs mysecutec commands on your behalf, so it uses your active environment (mysecutec use <env>) and login. If you're not authenticated, the command reminds you to run mysecutec login (the agent can also do this).
• The AGENTS.md written into the directory is the same guide documented under AI Agents; agents auto-discover it in the working directory.
• The default prompt asks the agent to confirm authentication and give a posture overview, and to ask before making any state changes.