CodeSignal measures employee skills from everyday work artifacts

What is changing?

CodeSignal made its Agentic Skills Intelligence platform generally available on October 28. Specialized agents examine real work artifacts from tools including GitHub, Salesforce, Gong, Cursor, Claude, ChatGPT, Workday, Greenhouse and Google Workspace. The company says the outputs are scored against structured rubrics grounded in industrial-organizational psychology, aiming to build evidence-based skills profiles instead of relying on titles, self-assessments or keyword inference.

Organizations can use these profiles to identify skills gaps across teams and roles, plan project staffing and track development. CodeSignal’s AI tutor, Cosmo, connects identified gaps to practice and coaching sessions tailored to an employee’s work context. An MCP server also lets teams query skills and jobs data from the AI assistant of their choice.

As assessment moves closer to real work, rubric transparency, data-access permissions, employee notice and attribution in jointly produced work remain critical. Organizations should not use automated scores alone for high-impact decisions such as promotion or dismissal; they should define coverage limits, human review and an appeal path.