Security audit
Security checks for vulnerabilities and agentic risk
This skill is only a Markdown instruction file, but it directs an agent toward broad autonomous legal work including automated court filing without clear user approval or human review.
Review before installing. The package does not appear to contain malware or executable code, but its instructions could steer an agent into sensitive legal actions. Use it only with explicit user confirmation for each external action and qualified legal review before any filing, demand letter, discovery action, or dispute submission.
The workflow explicitly describes automated filing through API-integrated court dockets without any warning, human approval requirement, or safety gating. Because legal filings can create immediate real-world consequences, this could enable unauthorized filings, fraudulent submissions, harassment, or irreversible procedural actions at scale.
The manifest description is extremely broad and lacks clear activation boundaries, covering autonomous advocacy, discovery, and dispute resolution across jurisdictions. In a high-impact legal domain, vague scope increases the chance of inappropriate triggering, overreach into unauthorized legal tasks, or use outside intended safeguards.
