dev-agent

Benign-to-moderately-risky overall. The skill footprint coherently matches its stated autonomous development purpose, leveraging MCP-integrated tools to fetch tasks, implement changes, and update stakeholders. The primary risks are operational: ensure least-privilege access, proper audit trails, and human-in-the-loop safeguards for high-impact changes. No evident credential harvesting, external exfiltration, or unverified binaries are described. Emphasize explicit user confirmations for complex tasks and strict access control/config validation to maintain safety in multi-repo environments.