Sysdig has made Sysdig AI Defense generally available, a product the company says watches every action an AI agent takes and uses an organization's policies to allow it, ask the user to confirm it, or block it before it completes. The company announced the launch on Nov. 4, 2026, and said the product is available for macOS and Windows.
According to Sysdig's announcement, agents across platforms such as Claude Code, Codex, GitHub Copilot and Cursor can be traced to an accountable person. The company says the product builds a live inventory of agents, harnesses, models, Model Context Protocol (MCP) servers, skills and plugins from runtime telemetry, including tools nobody approved.
Runtime security controls
Sysdig says its security controls judge each action by what an agent sets out to do, what it actually does and what its credentials can reach. A company blog post by Blair Howard says everything a customer's policies do not name runs as normal. For named actions, the post says, "Confirm pauses the call and asks the developer to approve it first. Block refuses it outright."
The press release says the product can stop unvetted MCP servers, skills and plugins before they run, and can block or hold risky credential reads and behavior drift for human review, with exception paths for developers. If an action gets past policy, Sysdig says the product kills the process. Policies start in alert mode and move to enforcement when a team is ready, and templates include rules mapped to the OWASP Top 10 for LLM Applications, according to the company.
Sysdig gives an example of how context changes the call. "Reading a credential to a test environment is noise, but reading the key that administers production is an incident," the press release says.
Continuous monitoring and audit records
Sysdig describes continuous monitoring in which AI Defense records every session as it happens and keeps a record at the kernel level that a compromised or rogue agent cannot tamper with. A live action graph pairs what an agent said it was doing with what it actually did and whose authority it acted under, according to the release. The company says customers get immutable records showing which policies were applied, when, and who signed off, and the blog post says records can be forwarded to a SIEM for security operations.
For incident response, Sysdig says its session replay lets teams trace tool calls alongside detections and the policies that fired.
The company says early data from the product shows that for every action an agent writes in its log, the kernel sees about seven processes doing the work. SecurityWeek, reporting on the launch, described it as five or more processes visible only to the kernel.
"An agent can misreport a tool call, but it can't rewrite its interactions with the system or the external world," said Loris Degioanni, Sysdig's founder and chief technology officer, in the release.
Chief executive Hatem Naguib said in the release that "every organization is now responsible for what its AI agents do." He added that agents "may start on the endpoint, but their blast radius extends well into the cloud."
Sysdig also says the product attributes every model call to an agent, team and person, including calls that never pass through an AI gateway, so leaders can track AI spending.
Availability and customers
Sysdig says this security solution uses the same runtime engine it built for cloud security, with kernel-level instrumentation rooted in Falco, the open source runtime detection project. The company says existing customers can contact their account teams, and others can request a demo or start a 14-day free trial. Sysdig did not publish pricing in the announcement.
The release includes a quote from Jameel Al-Aziz, chief technology officer and chief information security officer at Bastion, who said: "When an auditor asks what an agent did and who approved it, we have a trustworthy record."
The release also cites an IBM survey in which 77% of organizations said AI adoption is outpacing their governance capabilities.
Source: Sysdig, “Sysdig Launches Sysdig AI Defense, Bringing Runtime Security to Every Action AI Agents Take”
Additional sources: Sysdig, Blair Howard, “Introducing Sysdig AI Defense: See, stop, and prove what agents do”; Sysdig, “Sysdig AI Defense”; Sysdig, “Sysdig AI Defense Security”; SecurityWeek, Kevin Townsend, “Sysdig Adds Runtime Security and Governance for AI Agents”
