Topic
Security of service accounts, workload identities and AI agents that hold privileged access and are created dynamically in SaaS and cloud environments.
AppViewX's Agent Identity Security now finds browser-based and short-lived shadow AI agents and can terminate any agent and its sessions at runtime. It handles agents as identities to scope and revoke, and every claim about how well that works comes from AppViewX and one customer, ZoomInfo.
Reality
SailPoint says it will now stop bad access as it happens, citing up to 150 million AI agent interactions a day at an average Global 2000 company. Its own survey finds 15% of organizations can provision non-human access in real time, and removing agents' standing access depends on that capability.
Reality
SailPoint added AI agent discovery and automated access cleanup to its identity platform, aimed at the 79% of enterprises it says run agents in production. Existing customers can bring agent credentials under the platform that governs their staff, on SailPoint's evidence alone so far.
Reality
Nvidia is positioning its OpenShell runtime and Open Agent Safety Platform to contain AI agents, tying agent authority to independently proven control. The only public account is a single trade brief, so security teams can apply the principle to their own agents today while the product's containment claims wait for testing.
Reality
Rig Security raised $12 million in seed funding to separate an AI agent's actions from those of the employee whose identity it borrows. The standard way to stop a misbehaving agent, disabling its account, locks out the person as well.
Reality
Pillar Security CEO Ziv Karliner says OpenAI's sandbox escape shows AI agent limits must be enforced outside the model. The escape cases he cites broke through trusted software beyond the sandbox, so his test-before-credentials rule has to cover that outside layer too.
Reality
GitGuardian's 2026 State of Secrets Sprawl Report says commits identified as AI-assisted leak secrets at about twice the rate of human-written ones. Agent and MCP config files also keep plaintext keys on developer machines where commit scanning and code review never look.
Reality
Enterprise Cloud owners can now pull owner, scope, expiry and last-use data for SSH keys, PATs and app tokens as a CSV. The export stops at credentials GitHub issued, so secrets pasted into repositories stay out of scope.
Reality
Growth Equity at Goldman Sachs Alternatives has put $400m into Cyera at the valuation the company set in June, its third large round of 2026. What that cheque funds and what a security team can switch on next month are separate questions.
Reality
A zero-click RCE hit four AI coding agents through 925 hijacked plugins. A dev.to post pins the reach of it on keys that never expire. Its replacement is a grant scoped to one run and one task.
Reality
Hush Security searched public GitHub for the configuration filenames coding agents write, classified how each credential slot gets its value, and found a hardcoded literal in 12% of them, most of which match no vendor token format.
Reality
Martin Kuppinger of KuppingerCole argues role-based access cannot hold an agent that makes probabilistic, multi-step decisions, and wants per-session credentials plus logs recording what the agent was told to do.
Reality
Okta wants entitlements for AI agents and workloads evaluated continuously instead of recertified on a schedule, on the argument that approvals accumulate and access changes in the gaps between reviews.
Reality
The token was unrelated to the task the agent was given, and its blanket GraphQL permissions covered the volume-level backups too. GitGuardian counts 24,008 secrets in public MCP config files, 2,117 of them still valid.
Reality
The SANS 2026 survey puts 74 percent of businesses on standing credentials for autonomous AI, with no single control above 40 percent. The two MCP incidents most often cited as proof have different root causes.
Reality
Ping Identity executives split enterprise AI agents into four classes at their Austin conference. The two that live on endpoints have no identity of their own, so their actions authenticate as the employee who lent them the credential.
Reality
The new controls compare an agent's runtime behaviour against its authorized scope and can revoke the credentials underneath it. The 57% unmanaged-identity figure behind the pitch is Orchid's own, and neither publisher gives the method.
Reality
Reco says four in five AI tools in its telemetry run outside IT oversight, averaging 414 per 1,000 staff at smaller companies, which turns the cleanup into a question of who now owns each agent's credentials.
Reality
Pillar Security's Dor Sarig argues the non-human identity buildout answers who an agent is rather than what it does, and cites an internal agent any Slack message could trigger across a thousand private repositories.
Reality
Mitiga field CISO Brian Contos argues most enterprises cannot say what a live agent did after the fact, and the people who feel that first are the auditors and responders who have to produce a sequence.
Reality