CISO
Enterprise Architect
JUNE 30, 2026
A published flaw shows AI coding agents can approve one command and execute another, breaking the audit trail approval processes are meant to create.
Adversa AI researcher Omer Ben Simon published GuardFall on June 30, 2026, a class of shell-interpretation bypasses affecting open-source AI coding and computer-use agents. Testing found 10 of 11 surveyed agents (Hermes, opencode, Goose, Cline, Roo-Code, Aider, Plandex, Open Interpreter, OpenHands, and SWE-agent) vulnerable. The bypass exploits a mismatch between how approval guards inspect a proposed command and how Bash rewrites that command before execution, allowing decades-old shell tricks such as quote removal, IFS expansion, and command substitution to slip destructive commands past a filter that only reads the literal submitted text. Continue was the only tested agent found to substantially mitigate the issue. Adversa demonstrated end-to-end exploitation against the production Plandex binary. No CVE has been assigned, since the issue is a structural pattern across implementations rather than a single patchable flaw.
GOVERNANCE IMPLICATION
Approval layers built to review AI agent commands before execution assume the reviewed text and the executed text are the same artifact. GuardFall shows that assumption fails structurally in most tested open-source coding and computer-use agents, because Bash rewrites raw command text after a pattern-based guard has already approved it. Enterprises that rely on command approval logs as evidence of controlled agent behavior cannot treat those logs as proof of what actually ran. Any control environment that grants agents shell access and depends on approval-before-execution as a safeguard should treat this as an open gap until the agent's guard evaluates the post-expansion form of a command, not the submitted string.
SCENARIO
A development team enables auto-approval for an AI coding agent inside a CI pipeline, trusting a command-approval guard to block destructive shell operations. A poisoned dependency file contains an instruction that, once expanded by Bash, deletes credential files the guard never flagged because the submitted text looked benign. The approval log shows a clean review. The execution log shows something else.
THE GOVERNANCE QUESTION
If the command an approval process reviews is not provably the command that executes, what does an approval log actually prove?
CONTROL GAP
Command approval guards in most tested agents evaluate raw submitted text rather than the shell-expanded form Bash actually executes, so an approval decision does not verify the action that runs.
PRIMARY SOURCE
AI coding agents vulnerability: GuardFall shell injection
Omer Ben Simon, Adversa AI
June 30, 2026
Read the primary source ->(opens in new tab)CONTINUE READING
JULY 1, 2026
Agent SecuritySysdig's Threat Research Team documented JADEPUFFER, published July 1, 2026, describing what it assesses as the first documented case of agentic ransomware. A human operator provisioned infrastructure, selected the victim, and supplied MySQL credentials obtained through a prior compromise. From initial access, an LLM agent autonomously executed reconnaissance, credential harvesting, lateral movement, persistence, and destruction against a production database server, exploiting a Langflow remote code execution flaw (CVE-2025-3248) and a 2021 Nacos authentication bypass (CVE-2021-29441). Sysdig director of threat research Michael Clark told CyberScoop a human set up and pointed the operation while the agent executed the technical chain unsupervised. The agent diagnosed a failed login attempt and deployed a working fix in 31 seconds, and ran more than 600 distinct payloads before encrypting 1,342 database configuration records and deleting the originals.
JUNE 9, 2026
Agent SecurityAnthropic launched Claude Fable 5 and Claude Mythos 5 on June 9, 2026. Fable 5 is the first Mythos-class model released for general use. It includes safety classifiers that intercept queries in cybersecurity, biology and chemistry, and distillation categories, routing those queries to Claude Opus 4.8 instead. Anthropic reports the fallback occurs in fewer than 5% of sessions. The launch introduces a mandatory 30-day data retention requirement for all Fable 5 and Mythos 5 traffic on first- and third-party surfaces. Anthropic states the retained data will not be used for model training and will be deleted after 30 days in most cases.
MAY 18, 2026
Agent SecurityOn May 18, 2026, NIST published 'Summary Analysis of Responses to the Request for Information Regarding Security Considerations for AI Agents' (NIST Trustworthy and Responsible AI, report 800-5, authored by Riggs, Hamin, Perry, Edelman, and Cihon). The report summarizes stakeholder responses to the CAISI request for information (docket NIST-2025-0035). Commenters broadly agreed that AI agents present novel security threats that act as a barrier to adoption, and that while core cybersecurity principles still apply, they require adaptation for agents. Respondents identified roles for government including implementation guidance, information-sharing, and standards.