Add run_all cli + route

[cabch]

No description provided.

[github-actions]

This update introduces an automated process that runs multiple language model attacks using the given specifications, with results accessible through a new API endpoint. These enhancements include a command-line interface method and a JSON file designed to define default attack parameters, boosting the software’s ability to test various model vulnerabilities systematically.

Walkthrough

• New Feature: Added run_all subcommand to CLI for executing LLM attacks using JSON specifications.
• New Feature: Introduced /run_all API endpoint to facilitate running attacks programmatically, enhancing automation capabilities.
• Chore: Introduced default.json specifying attack configurations and models, providing a structured default approach for various attack tests.

_{Model: gpt-4o | Prompt Tokens: 1324 | Completion Tokens: 136}

[github-actions]

Here's a supportive code review enhanced by AI assistance. Please note that some observations are AI-generated suggestions rather than definitive conclusions, so trust your judgment in applying what feels most valuable. You remain the decision-maker in your development process—AI simply offers additional perspective to complement your expertise.

---

Always critique what AI says. Do not let AI replace YOUR I.
_{Model: anthropic--claude-4-sonnet | Prompt Tokens: 2744 | Completion Tokens: 1699}

[cabch]

modification done