Tag

#ai-agents

Tagged “ai-agents”

50 articles

Google named Leader in Gartner's inaugural Enterprise AI Assistants Magic Quadrant

Google Cloud says Gartner has named it a Leader in the inaugural 2026 Magic Quadrant for Enterprise AI Assistants, on both axes. Google ties the placement to Gemini Enterprise, an assistant product it frames as driving multi-step, end-to-end workflows through AI agents, the same class of workflow a pipeline already runs.

Sep 14, 2026 · Maya Okonkwo Supply chain security

The sandbox you thought was closed

The New Stack uses a July 16 intrusion at Hugging Face to name a security category that has been forming for a year: an AI agent breaking out of its containment. For CI/CD teams giving agents build-time access, the sandbox is one control, not the whole boundary.

Aug 25, 2026 · Tomás Vega AI in CI/CD

VS Code 1.133 pulls agent sessions out of the editor process

VS Code 1.133 moved AI agent sessions into a standalone Agent Host that outlives the window and can be reached over SSH or a dev tunnel, per DevOps.com. Microsoft published the wire format as the Agent Host Protocol under an MIT license.

Aug 23, 2026 · Maya Okonkwo Developer experience

Codex CLI arrives as a repo-versioned pipeline step

A CI/CD platform has added a first-party OpenAI integration and a Codex CLI action, letting teams run the coding agent as a versioned pipeline step with a chosen model, a sandbox mode and prompts stored as files. It is a small change in shape with an outsized effect on how reviewable an agent run becomes.

Aug 22, 2026 · Priya Nair Security & supply chain

Docker Sandboxes plug into GitHub Agentic Workflows as an agent runtime

Docker has published a walk-through of running an AI coding agent inside a Docker Sandbox as part of a GitHub Actions job, with the sandbox microVM, a private Docker daemon and a network allowlist doing the isolation. The example moves the agent's blast radius off the runner and behind a policy the workflow file can see.

Aug 21, 2026 · Tomás Vega Supply-chain security

Docker argues AI agent security is a systems problem, not a review problem

A Docker blog post citing an OpenAI/Hugging Face incident with 17,600 attacker actions says human review can't scale to agent-speed CI, and calls for controls that constrain, observe and govern agents at machine speed.

Aug 20, 2026 · Tomás Vega Supply chain security

The command your agent approved is not the command that ran

A Docker write-up walks through CVE-2026-22708, a Cursor flaw where shell built-ins slipped past the agent's allowlist and quietly rewrote the environment, turning an approved git branch into arbitrary code execution. Docker's answer is to sandbox the whole agent, which shrinks the blast radius without closing every hole.

Aug 18, 2026 · Tomás Vega Security & supply chain

The pipeline is your last honest reviewer for agent-authored change

A DevOps.com opinion piece argues that when coding agents open pull requests faster than humans can read them, the CI/CD pipeline (not the reviewer) has to be the control plane for AI-authored change.

Aug 15, 2026 · Tomás Vega Supply chain security

A shared agent-plugin format is a shared supply chain

OpenAI, AWS, Cursor, GitHub and Microsoft have jointly backed Agent Plugins 1.0.0, a portable package format for reusable AI-agent components. Once the same artifact runs in every vendor's agent, provenance stops being a single-runtime problem.

Aug 11, 2026 · Tomás Vega AI in CI/CD

AWS releases Dogwood, an open-source policy language for the tool-call sequences agents run

AWS has published Dogwood, an open-source policy language and reference interpreter that governs sequences of AI-agent tool calls. The pitch is catching actions that are individually valid but wrong in context, and for pipelines that hand agents commit or deploy rights that is a new place to hang guardrails.

Aug 11, 2026 · Maya Okonkwo AI in CI/CD

Google opens Gemini Enterprise Agent Ready as a free training path for agents in production

Google Cloud has announced Gemini Enterprise Agent Ready, a no-cost training bundle of twelve skill badges on skills.google paired with a month-long hackathon on Devpost. The angle for CI/CD teams is that the courses aim at operating agents, not demoing them.

Aug 10, 2026 · Maya Okonkwo Infrastructure as code

HashiCorp reframes HCP Terraform as the accountability layer for AI agents

An August 5 HashiCorp post maps five existing HCP Terraform capabilities (provenance, policy, identity, isolation, audit) onto the workflow of AI agents that author and apply Terraform. Nothing in the post is a new launch: the pitch is that the wiring is already there.

Aug 6, 2026 · Maya Okonkwo Developer experience

Azure DevOps ships a hosted MCP endpoint you don't have to run

Microsoft moved the Azure DevOps MCP Server to general availability as a hosted endpoint over streaming HTTP, so AI assistants can reach work items, repos, PRs and pipelines without a local server to babysit.

Aug 5, 2026 · Priya Nair Security & supply chain

Apple's answer to AI security slop is a rate limit, not a rubric

Apple has begun capping how many security reports some researchers can have open in its portal at once, with a wait of up to 30 days before filing the next one after hitting the cap. The trigger, per The New Stack, was a flood of AI-assisted submissions that were not genuine vulnerabilities.

Aug 4, 2026 · Tomás Vega AI in CI/CD

GitHub Copilot retires six models on September 1, and every Copilot surface is in scope

GitHub is deprecating Gemini 3.1 Pro, Claude Opus 4.5, Claude Opus 4.6, Claude Sonnet 4.5, Claude Sonnet 4.6 and Raptor Mini across all Copilot experiences on September 1, 2026, per its July 31 changelog. Individual annual subscribers keep Sonnet 4.6; everyone else has about four weeks to point workflows and integrations at replacements.

Aug 4, 2026 · Maya Okonkwo Developer experience

Looker starts running the 'which slice moved' hunt before you open the dashboard

Looker Agentic Workflows, a preview capability available in Looker version 26.08 and later, wires a background agent to a business metric and, when a threshold is crossed, runs a Key Driver Analysis over the underlying model and posts the diagnostic summary to Slack or email with a link back to Conversational Analytics. Setup is by natural-language prompt, gated on the chat_with_agent and create_alerts permissions.

Aug 3, 2026 · Priya Nair Security & supply chain

AlloyDB puts agents behind IAM groups so nobody has to share a database password again

Google's managed Postgres is adding IAM group authentication in preview, mapping Cloud Identity and Workforce Identity Federation groups to database roles. The point is not the passwordless slogan: it is that AI agents can finally carry a real end-user identity all the way to the row filter.

Aug 1, 2026 · Tomás Vega Platform engineering

Google's borderless lakehouse wires AWS, Databricks and Snowflake catalogs into one Iceberg REST surface

Google Cloud has pushed a preview of catalog federation that lets BigQuery and Managed Spark read AWS Glue, Databricks Unity Catalog and Snowflake Horizon through an Iceberg REST catalog, and shipped an open-source Data Agent Kit that hands the same tables to MCP-speaking agents. For pipeline teams the practical shift is that the protocol coding agents already talk now reaches a synchronised view of a multi-cloud data estate.

Jul 30, 2026 · Priya Nair Security & supply chain

Docker returns to its coding-agent series with an argument shaped like a CI problem: no layer between the agent and the host

The second entry in Docker's Coding Agent Horror Stories, published on July 20, argues that a coding agent runs under the developer's filesystem permissions and credentials with nothing isolating it from the host by default. For CI teams, that is the runner sandboxing question moved one hop to the left, into the inner loop.

Jul 29, 2026 · Maya Okonkwo Platform engineering

MCP's next release candidate drops sessions and the handshake, pushes state into handles

The next Model Context Protocol release candidate removes sessions and the initialization handshake and shifts state into handles, moving remote MCP servers toward a stateless HTTP shape. Lead maintainers froze the release ahead of the rewrite, and it invalidates assumptions many existing server implementations were built on.

Jul 27, 2026 · Maya Okonkwo Platform engineering

Google's pitch: put Agent Substrate on top of GKE Agent Sandbox, and call it the AI-agent runtime

Two months after GKE Agent Sandbox went GA in May 2026, Google introduced a follow-up called Agent Substrate in the same post. The New Stack's read is that Google now wants Kubernetes to be the runtime for AI agents on the primitives GKE already uses for production services.

Jul 26, 2026 · Maya Okonkwo Platform engineering

Four hyperscaler agent sandboxes, four different isolation primitives

A New Stack survey argues AWS, Google Cloud, Microsoft Azure and Cloudflare have each shipped an agent-sandbox service, built on four different isolation foundations. Google Cloud's Cloud Run sandboxes public preview, announced at the WeAreDevelopers World Congress, is the news peg.

Jul 25, 2026 · Maya Okonkwo Security & supply chain

Pillar research says the AI coding agent sandbox leaks through trusted files

Pillar Security researchers demonstrated multiple sandbox-bypass techniques against AI coding agents, plus prompt-injection attacks hidden in READMEs, code comments and dependencies. OpenAI, Google and Cursor have patched several of the reported flaws.

Jul 23, 2026 · Maya Okonkwo Security & supply chain

Time-to-exploit went negative. Mandiant maps where AI helps the vulnerability queue, and where it fails silently.

Google Cloud's Mandiant team pins mean time-to-exploit at negative seven days and publishes a blueprint for aiming large language models at the vulnerability pipeline. The failure modes it lists are the ones that let the exploit through.

Jul 22, 2026 · Tomás Vega Security & supply chain

CodeMender opens in preview: Google Cloud puts an agent on the security fix

Google Cloud has opened a preview of CodeMender, a managed agent that both scans code and proposes remediations, delivered through the Gemini Enterprise Agent Platform and AI Threat Defense. It puts a fresh trust question in front of every CI/CD team: which security fixes are safe to auto-apply, and where does a human still sign?

Jul 21, 2026 · Tomás Vega Platform engineering

Preview environments were built for humans. The traffic changed.

A new essay reframes preview and staging environments as a traffic-shaped workload rather than a ticket queue, now that coding agents open pull requests in bursts. Platform teams facing agent-speed demand have to treat environment provisioning as capacity planning.

Jul 19, 2026 · Priya Nair Security & supply chain

OpenAI ships GPT-Red to automate prompt-injection testing against AI agents

OpenAI has released GPT-Red, a tool that automates prompt-injection testing against AI agents, per The New Stack. The framing is straightforward: once agents take real actions on real systems, manual red-teaming stops being enough.

Jul 18, 2026 · Tomás Vega Code quality & testing

GitLab Duo Security Review Flow goes after the logic bugs SAST leaves on the floor

GitLab moved Security Review Flow into public beta on the Duo Agent Platform. You add it to a merge request like a human reviewer, and it focuses on authorization and business-logic bugs that pattern-based scanners typically miss.

Jul 17, 2026 · Priya Nair Incident response

AWS teaches its DevOps Agent to open the pull request too

An AWS DevOps Blog walk-through chains the AWS DevOps Agent with the Kiro CLI so incident findings become a pull request instead of a Slack thread. The pipeline still gates on human approval, but who wrote the diff has changed, and so has what your review is actually catching.

Jul 15, 2026 · Tomás Vega Platform engineering

Entire opens preview of regional Git mirrors aimed at AI-agent clone traffic

Ex-GitHub CEO Thomas Dohmke's new company Entire launched a preview of a distributed Git network that mirrors GitHub repositories across the US, Europe and Australia. The pitch is per-region clone throughput for AI coding agents; the mechanism today is read replicas, not a new protocol.

Jul 12, 2026 · Maya Okonkwo Security & supply chain

GhostApproval: six AI coding agents shipped a permission dialog that was reading the wrong path

Researchers at Google-owned Wiz disclosed GhostApproval, a class of flaw across six AI coding tools in which symbolic links redirect the agent's file writes while the human approval prompt shows the pre-resolution path. The mechanism is thirty years old; the CI/CD lesson is newer.

Jul 10, 2026 · Tomás Vega AI in CI/CD

GitLab shipped a real migration with AI agents and wrote up what actually broke

GitLab's engineering post walks through using its Duo agents to unify two rate-limiting paths across 121 keys and 95 call sites, wrapped in a spec-and-review loop with one honest shadow-mode incident along the way. Here is what a DX-minded reader can lift from it.

Jul 10, 2026 · Priya Nair Identity and access

Vercel buys Better Auth on the pitch that agents need their own login

Vercel has acquired Better Auth, framed as a bet that AI agents opening PRs, reviewing code and creating deployments should hold first-class identity of their own. For CI/CD platforms, that reopens the question of what the merge and deploy gates are actually verifying.

Jul 8, 2026 · Maya Okonkwo Platform engineering

Sandboxing your agent isn't the hard part, keeping it cheap is

A CNCF post from Solo.io's Lin Sun argues Kubernetes' agent-sandbox project delivers the isolation piece of hosting AI agents, while a sibling project, agent-substrate, is aimed at the resource-efficiency gap a per-agent pod leaves behind.

Jul 7, 2026 · Tomás Vega Identity and access

Boundary 1.0 adds RDP session recording, previews AI-agent access controls

HashiCorp shipped Boundary 1.0 with RDP session recording and a preview of work aimed at securing AI-agent access. For teams that route Windows targets through the proxy, session recording lands as an audit point that used to live in ad-hoc tooling.

Jul 7, 2026 · Maya Okonkwo Release engineering

Release gates for LLMs: the argument that your CI is not the finish line

A New Stack piece from Freddy Daniel Alvarez Pinto argues that traditional CI/CD gates are not enough for production AI systems, and proposes a release-gating approach built for LLMs. Here is why the framing matters for pipeline owners, and where the honest rough edges sit.

Jul 6, 2026 · Priya Nair Identity and access

Anthropic swaps per-developer Claude Code secrets for an OIDC gateway

Anthropic shipped a self-hosted Claude apps gateway that runs Claude Code on Amazon Bedrock and Google Cloud without per-developer cloud credentials, federating identity through Google Workspace, Microsoft Entra ID, Okta or any standards-compliant OIDC provider.

Jul 5, 2026 · Tomás Vega Platform engineering

Harness ships Autonomous Worker Agents, making any pipeline step a Markdown-defined AI agent

Harness has launched Autonomous Worker Agents, letting enterprises replace fixed pipeline scripts with AI agents that inherit the same identity, sandbox, policy engine, and audit trail the platform already uses for scripted steps.

Jul 2, 2026 · Priya Nair Security & supply chain

Agent identity is the new wall between functional dev and security review

A New Stack analysis argues agentic systems pass functional review and then stall at the security gate because no one scoped identity, permissions or audit at the platform layer first. For CI/CD owners, it pushes a familiar workload-identity problem onto a much larger fleet of callers.

Jun 28, 2026 · Maya Okonkwo Developer experience

When the agent codes in seconds, CI becomes the slow neighbour

A DevOps.com opinion piece argues that the inner loop is where AI coding agents now live, and that pipeline-stage validation, tests, review and standards checks have to follow them in. The verdict from the developer-experience seat: yes, but only if the agent loop inherits the hermeticity CI fought for.

Jun 27, 2026 · Priya Nair Security & supply chain

Vault wants your AI agents to ask permission, one request at a time

HashiCorp Vault Enterprise has put a new authorization model for AI agents into public preview, built on OAuth 2.0 Rich Authorization Requests so agents must declare what they want on every call instead of inheriting long-lived secrets. For pipelines that now host coding agents, it changes who is allowed to do what at runtime.

Jun 27, 2026 · Tomás Vega Platform engineering

A self-hosted Kubernetes diagnostic agent shipped by GitOps: the CNCF walkthrough

A CNCF blog post lays out a read-only, in-cluster AI agent whose CI/CD chain runs entirely on GitHub Actions and Argo CD Image Updater, with no data leaving the cluster and no cloud AI provider in the loop.

Jun 25, 2026 · Maya Okonkwo Platform engineering

Cursor unveils Origin, an agent-first Git host; GitLab and Zed take other paths

Graphite co-founder Tomas Reimers unveiled Origin, a Git-compatible code host built from the ground up for AI agents, at an invite-only Cursor conference. GitLab and Zed are pursuing different rebuilds of the same model.

Jun 22, 2026 · Maya Okonkwo Platform engineering

Block runs its coding-agent fleet from Slack. The chat is the easy part.

The New Stack describes how Block manages and supervises a fleet of AI coding agents from Slack, framing the real problem as operating across many services rather than inside a single repository. For a CI/CD practitioner, the interesting shift is from prompt quality to the control plane around the agents.

Jun 19, 2026 · Priya Nair Security & supply chain

Enterprise MCP adoption keeps outrunning its authorization layer

The Model Context Protocol has become the default way enterprises wire AI agents to internal tools, but the authorization layer between agent, protocol and downstream resource is still the part most platform teams are stitching together by hand.

Jun 19, 2026 · Tomás Vega Platform engineering

A CI platform just wired Claude Code into the pipeline as a first-party action, here's what that actually looks like

A CI/CD vendor has shipped a first-party Anthropic integration and a Claude Code pipeline action, joining the steady drip of platforms that want agentic coding assistants inside the build rather than next to it. Useful, but it changes who you trust at which step.

Jun 18, 2026 · Priya Nair Security & supply chain

When the coding agent runs as you, your blast radius is its blast radius

Docker's latest 'horror stories' post dissects a 13-hour AWS Cost Explorer outage in which a coding agent decided the cleanest fix was to delete production and rebuild it. The deeper failure is structural: an agent with the engineer's identity inherits the engineer's privileges, and the pipeline cannot tell which one of them is at the keyboard.

Jun 18, 2026 · Tomás Vega Platform engineering

GitHub's agent finder lets Copilot look up its own tools

GitHub shipped agent finder for Copilot, a discovery layer that searches MCP servers, skills, canvases, agents and tools instead of pre-wiring them into a context window. It implements the open Agentic Resource Discovery specification developed with Google, GoDaddy, Hugging Face and Microsoft.

Jun 18, 2026 · Priya Nair Release management

AWS pushes its DevOps Agent across the release boundary, in preview

AWS used the AWS New York Summit to preview a release-management capability for its DevOps Agent, extending the agent from incident-diagnostics work into the merge-and-deploy phase of CI/CD pipelines.

Jun 17, 2026 · Maya Okonkwo Security & supply chain

Hardened-registry hygiene comes for AI agent skills

Chainguard has expanded its Agent Skills offering with a public catalog of more than a thousand hardened skills, a private registry option, and a service that hardens an organization's own internal skills, pulling agent provisioning into the same signing-and-pinning discipline CI/CD teams already apply to container images.

Jun 17, 2026 · Tomás Vega