Critikal is an offensive security research tool. If you find a vulnerability in Critikal itself (not in the targets it audits), please report it privately.

• Preferred: open a GitHub security advisory.
• Alternative: email the maintainers (see pyproject.toml authors field).

Please include:

• A description of the issue and its impact.
• Steps to reproduce (or a PoC if possible).
• The Critikal version / commit you tested.
• Whether the issue affects a user's host (e.g., sandbox escape, credential exfiltration from .env) or only the quality of findings.

We aim to acknowledge reports within 72 hours and ship a fix or mitigation within 30 days for high-impact issues. Please do not file public issues or pull requests for undisclosed vulnerabilities.

In scope:

• Sandbox escapes from the bash, sandbox_run, or deploy_contract tools.
• Path traversal or arbitrary file write via file_write / file_edit.
• Credential leakage in logs, reports, or tool output.
• Prompt-injection vectors that cause Critikal to exfiltrate secrets or run unintended destructive shell commands against the host.
• Dependency vulnerabilities in pinned runtime deps.

Out of scope:

• False positives / false negatives in audit findings — file a normal issue.
• Issues in the target repositories Critikal analyzes.
• Model-provider downtime or API errors.

Critikal executes LLM-generated shell commands and Foundry tests. Run it in a container or a throwaway VM when analysing code you don't trust. Never point it at production credentials, private keys, or mainnet RPC URLs that can sign transactions.