
Stale job access lingers because nobody owns removal
The author says role transfers should trigger a deletion review, with named owners and expiry dates on extra rights.
24 stories, newest first · one card per story

The author says role transfers should trigger a deletion review, with named owners and expiry dates on extra rights.
On AgentDojo, tau-bench and ToolEmu, authors report 78.3% fewer tools per role and a 15.9% blast radius.
A deterministic loop checks each call in a median 5.7 microseconds, and flagged calls go to a reasoning guardian.

The piece treats MCP hosts like database drivers and lists controls from input checks to human approval of high-risk actions.
For 29.8–33.3% of Microsoft Graph tasks, minimal sufficient grants form an antichain with no smaller choice.

Vince Stoffer says teams should watch shadow AI, least privilege, change audits, and network use of AI.

The piece says education, sandboxes, and committees do not govern credentials, data access, or orphaned workflows after they go live.

Discovery now covers 60 local agent types, and risk signals can revoke access or stop an agent.

Dynamic Permissions start at zero and last only for a task; Guardian Agents follow human playbooks.
SANS argues defenders should build small least-privilege AI agents rather than caging general-purpose ones, citing Check Point's tool-call deviation detection.

Token Security explains how to stop AI agents from using credentials outside their assigned tasks.

Comparison of twelve just-in-time access tools across PAM, cloud-native, infrastructure, and governance lanes, with Microsoft Entra PIM as bundled baseline.

Rapid7 urges identity, least-privilege, and telemetry controls for autonomous AI agent-to-agent communication.

Unreleased Gemini Desktop Full Access could let the assistant control files, apps, and network activity on a Mac.

Guide proposes an IAM framework for AI agents, treating each agent as a governed non-human identity with scoped, expiring, auditable access.
Huntress outlines least-privilege steps to limit blast radius from excess admin rights in Microsoft 365.

Oasis Security executive argues enterprises need 'least agency' controls for AI agents, citing a Cursor agent's nine-second production database deletion and Claude Desktop flaw.

Gremlin security officer Frederic Bull argues AI has eroded the attacker-defender skill asymmetry while least-privilege controls remain essential for securing AI agents.

Cloudflare launches per-Worker granular access controls with four roles, enabling least-privilege access for teammates, AI agents, and CI/CD pipelines.

2026 CIEM guide ranks Wiz, Prisma Cloud, Okta, Entra Permissions Management and specialists Sonrai, Britive, Tenable/Ermetic for cloud entitlement right-sizing.

A 2026 buyer's guide compares 12 endpoint privilege management tools, ranking CyberArk and BeyondTrust as enterprise leaders.

A 2026 scorecard ranks ten endpoint privilege management tools, led by BeyondTrust, ThreatLocker and Delinea for elevation, coverage and policy depth.

Orchid's CEO recommends inventorying five attributes of AI agent credentials to rein in over-privileged OAuth tokens and service accounts.

MCP servers holding AI agent credentials risk secret exposure via plaintext configs, credential sprawl, prompt injection, and over-permissioning; mitigations include centralization and least privilege.
Today 1.5K readers · 2.4K views · 754 reads · all time 189K views