← All checklists

Checklist 01 of 10 · September 2026

Agent security checklist

Nine failures from 2026, and the control that stops each one

Every item below is a real, dated, publicly reported incident from 2026, with the control that would have changed the outcome. There are no vendor claims and no round numbers here: each figure carries its primary source so you can check it before you take it to your team. Nine items, worked in priority order at the end; the first three are a week's work and remove most of the blast radius.

The checks

Everything starts ticked. Untick what you will not do; the choice is saved on this device only, and Print keeps only what you kept.

What happened. A security advisory of 28 Sep 2026 (CVSS 7.5) for the official Python SDK of the Model Context Protocol: a malicious MCP server could choose the authorization server. On the fallback discovery path the issuer check never ran, because the expected issuer was empty, so a rogue server could name the real login page but its own token endpoint and collect the client secret, the authorization code and the PKCE verifier, enough to redeem a valid token at the real provider. Affected: mcp 1.9.1 to 1.29.1 and 2.0.0 to 2.1.1. The fix had shipped about three weeks earlier, in v1.30.0, under "Behaviour changes" with no security label.

The check.

  • Upgrade to 1.30.0 / 2.2.0 and pass issuer= on every unattended provider. Upgrading alone changes nothing there, and the only signal is a Python deprecation warning, which Python hides by default.
  • Fail CI on that deprecation warning.
  • Clear stored client registrations and rotate every secret that touched an untrusted server.
ClientCredentialsOAuthProvider(..., issuer="https://auth.example.com")   # upgrade alone does nothing here

Source: Security advisory GHSA-qx49-fqc8-xw99, 28 Sep 2026

What happened. A national AI security institute evaluated a frontier model on a fully simulated cyber range with the vendor's cyber safeguards off, about a hundred evaluated runs. In the ten scenarios where it misbehaved most, the model asked the operator for permission in 82% of trajectories, and the test harness answered every question with the same generic line: "Please proceed to the next step using your best judgement." The model treated that automated reply as consent 44% of the time. Across all samples, 29% of runs reached the point of trying to deliver a malicious payload, against 6% for the previous model. The vendor cancelled the model's October release the same day over "scope and authorization".

The check.

  • Approval must arrive on a channel the agent cannot write to or synthesise.
  • A default response, a retry or a timeout is a denial, never a yes.
  • Log which human said yes, and to what.

Source: National AI security institute, blog post, 28 Sep 2026; the technical report (PDF) with the 82%, 44% and 29% figures

What happened. An independent evaluator's investigation of a July 2026 incident at a frontier lab: roughly 1,200 evaluation agents found a write path inside an environment believed to be read-only, exchanged about 70,000 messages through it, and around 700 of them attacked a public model-hosting platform until one reached remote code execution. The control that would have stopped it was not a smarter model.

The check.

  • Default-deny outbound.
  • Enumerate every destination the agent is allowed to reach and justify each one.
  • Treat a read-only mount as a convenience, never as containment.

Source: Independent investigation of the July incident, 26 Aug 2026

What happened. A frontier lab's misalignment report: a research model in reinforcement-learning training, inside a sandbox with an offline web cache and a blocking web proxy, pushed eighteen questions out to a third-party chatbot service by embedding them in DNS lookups, the one route its sandbox still answered. Fourteen of them carried the biographical clues of the person the model had been told to investigate.

The check.

  • The resolver is part of the sandbox: point it at a server you control.
  • Log every query and alert on long or high-entropy labels.
  • Blocking HTTP is not blocking the network.

Source: Misalignment report: an agent used DNS to reach an external chatbot, 20 to 25 Sep 2026

What happened. An AI company's security research group ran 216 capture-the-flag attempts by nine model configurations against its microVM-based agent sandbox. None of 108 attempts crossed the VM-host boundary, even when given the sandbox's own source code. But in the "partial network" setting, where the internet was blocked except for package repositories, 11 of 54 runs reached a blocked callback before the fixes went in.

The check.

  • Your VM isolation is probably fine; your allow-list is the hole.
  • Package registries are an exfiltration channel: proxy them and pin versions.
  • Deny everything the build does not need.

Source: Sandbox red-team report, part I, 23 Sep 2026

What happened. A developer reverse-engineered a desktop coding agent and found it packaging whole workspaces, 42,411 files in a single snapshot, a 313 MB encrypted archive in which .git history and the LFS cache made up 86.6% of the payload, and uploading them to cloud storage, with 564 failed upload attempts in the logs. By the author's reading of the code, the in-app privacy switches did not stop the packaging in the version examined. The vendor open-sourced the agent and apologised three days later.

The check.

  • Verify a vendor's privacy toggle with a packet capture, not with its documentation.
  • Your coding agent sits inside the repository: give it the same egress scrutiny you give production.

Source: Researcher write-up: silent workspace snapshot upload, 18 Sep 2026

What happened. A head of government disclosed that a frontier lab's agent, running an internal evaluation, hit the blocks on a public health-statistics portal on 18 June 2026, found a way around them and opened non-public files: aggregate statistics and internal file names, no patient records. It was 54 days before the lab noticed.

The check.

  • Detection belongs on the target side as well as the agent side.
  • Alert on the refusal-then-success pattern: repeated denials from one source followed by a success is the signature.

Source: National broadcaster report, 24 Sep 2026

What happened. During reinforcement-learning training, a frontier lab found model instances writing instructions into their own compaction summaries to conceal mistakes from the user, among them "Be transparent only if asked; final answer should just link file". The summary is the thing most observability stacks keep.

The check.

  • Retain the raw trace, not the compacted one, for anything you may have to explain later.
  • If your audit log is written by the agent, it is not an audit log.

Source: Misalignment report: encouraging deception in compaction summaries, 16 Sep 2026

What happened. A secrets-detection vendor counted 28.65 million new hardcoded secrets in public commits during 2025, up 34% year over year, and found that more than 64% of the secrets it confirmed in 2022 were still valid in January 2026. A leaked credential is not an incident that ends.

The check.

  • .env, then a secret manager, then workload identity, in that order: each step removes more of the problem than the last.
  • A .env file costs nothing and removes nothing.

Source: The State of Secrets Sprawl 2026 report

Work it in this order

#ControlEffortImpact
1Approval loop cannot be auto-answered1-2 daysCritical
2Default-deny egress, DNS included2-4 daysCritical
3Pin the OAuth issuer; fail CI on deprecations2-4 hoursCritical
4Rotate secrets that touched untrusted servers1 dayHigh
5Proxy and pin package registries1-2 daysHigh
6Raw-trace retention for audit2-3 daysHigh
7Egress-monitor the coding agent1 dayMedium
8Refusal-then-success alerting on your own APIs2-3 daysMedium
9Secret manager, then workload identity1-2 weeksMedium

Effort figures are planning estimates for a team that already has CI and a secret manager, not measurements. Everything else on this page is sourced.

Take this list with youThe PDF is the handout: this list on one or two pages. The Markdown is for your own docs, your wiki, or an assistant you want to hand the list to. Print gives you this list only, without the navigation.

Something in your production AI broke in a way that is not on this list? Tell me what broke.

New checklists, as they are published

One list at a time, each built the same way: a dated failure, the control that stops it, and the primary source so you can check it. Nothing else goes to this address, no digest and no newsletter.