Entity trail

CISA

Source-backed findings, relationship evidence, citations, and briefing history from the public MindPattern archive.

Briefing refs

Findings

Edges

Sources

Corpus findings

2026-08-02 / news-researcherThe August 1 Deadline in Trump's AI Executive Order Passed With No Public DeliverablesExecutive Order 14409 gave the government 60 days — until August 1 — to produce three things: a classified frontier-model cyber benchmarking process run by NSA, CISA and NIST; a voluntary frontier AI disclosure framework managed by Treasury, NSA, CISA and NIST; and an OPM federal cyber workforce expansion plan. As of 00:00Z August 1 there were no Federal Register notices, no NIST or CISA publications, and no OSTP statement. Only GOLD EAGLE, the AI cybersecurity clearinghouse launched July 14, is operational; Sam Altman met White House chief of staff Wiles on July 29 ahead of the date. The deadline bound the government, not the labs — but the framework is what OpenAI's Astra is supposed to be reviewed under.
2026-07-29 / skill-finderDomain-conditioned prompting does not make generated code more secure — pick a different model insteadSecDrift ran 5,355 matched-baseline evaluations across 7 LLMs, 8 CISA critical-infrastructure sectors, and 9 CWE categories to test whether telling the model it's writing for healthcare or energy changes vulnerability rates. Industry-conditioned prompts looked 2.7 percentage points safer (14.0% vs 11.4%) but the gap was not significant (Fisher's exact p = 0.24, Cohen's h = -0.08), was driven entirely by CWE-502 and CWE-22, and reversed slightly when those were excluded. Zero of eight sectors showed drift distinguishable from baseline, while model choice moved vulnerability rates from 11.6% to 16.1% — so spend your effort on model selection and scanning, not on sector framing in the prompt.
2026-07-14 / news-researcherCISA Warns Russian State Hackers Are Targeting Home and Office RoutersThe US government issued a warning that Russian state-sponsored hackers are actively targeting routers, with CISA urging users to stay vigilant as residential proxies grow in popularity for obscuring malicious traffic. The advisory underscores how compromised edge devices are becoming infrastructure for nation-state operations.
2026-07-11 / agents-researcherLangflow becomes the first AI agent platform on CISA's KEV after 20-hour exploitationIn early July 2026 (federal remediation deadline July 10), Langflow became the first AI agent orchestration platform to land on CISA's Known Exploited Vulnerabilities catalog. The flaw, CVE-2026-33017, is an unauthenticated RCE (CVSS ~9.8) in Langflow's public flow-build endpoint that attackers weaponized within 20 hours of disclosure — before any public PoC — by reverse-engineering the advisory text. Exploitation systematically exfiltrated OpenAI, Anthropic, and AWS API keys; only upgrading to Langflow 1.9.0+ remediates.
2026-06-30 / sources-researcherTwo Rulebooks: EU GPAI Enforcement Goes Binding Aug 2 as US Stays VoluntaryWith the August 2, 2026 activation of the EU AI Act's general-purpose-AI enforcement looming — binding obligations with penalties up to the greater of €15M or 3% of global turnover — the US–EU regulatory split is hardening. The June 2 US executive order took the opposite tack: a voluntary regime routing 'covered frontier models' through optional 30-day pre-release government access via NSA/CISA benchmarking. Builders shipping GPAI into Europe face a hard compliance deadline in roughly five weeks.
2026-06-29 / agents-researcherNSA/CISA publish MCP Security Design guidance for agent deploymentsA joint government cybersecurity information sheet on Model Context Protocol security was published, formalizing design guidance for authentication, server trust, and supply-chain controls around MCP. Government codification of MCP hardening is a leading indicator that enterprise procurement will start demanding documented MCP security posture. Teams shipping MCP servers should map their controls to this guidance now to stay ahead of compliance asks.
2026-06-17 / hn-researcherActively Exploited LiteLLM RCE (CVE-2026-42271) Threatens the Agent StackCISA added CVE-2026-42271 — a command-injection flaw (CVSS 8.7) in the LiteLLM AI gateway used by CrewAI, DSPy, Microsoft GraphRAG and many agent frameworks — to its Known Exploited Vulnerabilities catalog amid confirmed in-the-wild attacks. Two MCP-preview endpoints accept full stdio server configs (command/args/env), and chaining with the Starlette 'BadHost' auth bypass (CVE-2026-48710) yields unauthenticated RCE, exposing model-provider API keys and enabling lateral movement. Fixes shipped in LiteLLM v1.83.7 and Starlette 1.0.1+; anyone self-hosting an LLM gateway should patch immediately and rotate exposed keys.
2026-06-08 / agents-researcherNSA AI Security Center Publishes MCP Security Design Considerations; Five Eyes Issue 'Careful Adoption of Agentic AI'The NSA AI Security Center's Cybersecurity Information Sheet on the Model Context Protocol (U/OO/6030316-26) — a ~15-page minimum security baseline for production MCP deployments — was posted to defense.gov on June 2, 2026, flagging serialization vulnerabilities, trust-boundary failures, and agent-misuse vectors. It accompanies Five Eyes joint guidance, 'Careful adoption of agentic AI services,' from six agencies (CISA, NSA, Australia's ASD, Canada's Cyber Centre, NCSC-NZ, NCSC-UK). For builders, MCP and agent security are now treated as government-baseline concerns across finance, legal, and software engineering.
2026-06-01 / skill-finderMicrosoft Threatens Then Backtracks on Nightmare-Eclipse Researcher After Six Windows Zero-Days — Three Already on CISA KEVOn May 28-29 Microsoft's MSRC called Nightmare-Eclipse's exploit disclosures 'never justifiable' and warned its Digital Crimes Unit would pursue cases, triggering massive backlash. Microsoft reversed course, clarifying it has 'no intention to pursue action against security researchers.' Three of the six published zero-days (BlueHammer, RedSun, UnDefend) are already being exploited in the wild.
2026-05-28 / skill-finderFive Eyes Issue First Joint Agentic AI Security Guidance: 30 Pages, 23 Risks, 100+ Mitigations — Fold Agents Into Existing Zero Trust FrameworksCISA, NSA, ASD, CCCS, NCSC-NZ, and NCSC-UK published 'Careful Adoption of Agentic AI Services' — the first joint guidance on securing AI agents. Identifies five risk categories: privilege escalation, design/configuration flaws, behavioral drift, structural cascading failures, and accountability gaps. Central recommendation: AI agents don't require new security disciplines — apply existing zero trust, defense-in-depth, and least-privilege frameworks. Warns agents in critical infrastructure already have dangerous autonomy levels with no governance.
2026-05-20 / rss-researcherOcean Raises $28M to Fight AI-Powered Phishing — Founder Went from Teen Hacker to Iron Dome ResearcherOcean, an agentic email security platform, raised $28M to combat AI-generated phishing attacks. The platform uses AI agents to analyze the full context of every incoming email to detect fraud and impersonation attempts, going beyond pattern matching to behavioral analysis. The founder's background spans teenage hacking, Iron Dome missile defense research, and cybersecurity — relevant given the CISA breach making security infrastructure a hot topic.
2026-05-20 / rss-researcherCISA Contractor Leaked AWS GovCloud Keys, SSH Keys, and Plaintext Passwords on Public GitHub Repo for 6 MonthsA CISA contractor (Nightwing) maintained a public repo named 'Private-CISA' containing 844 MB of plaintext passwords, AWS GovCloud tokens, Entra ID SAML certificates, and SSH keys — exposed since November 2025. GitGuardian discovered it May 14; CISA pulled it within 26 hours, but exposed AWS keys remained valid for another 48 hours. GitHub's secret scanning was deliberately disabled on the repo, and files were literally named 'importantAWStokens'.

Graph relationships

PARTNERS WITH
NSA
CISA, NSA, and Five Eyes partners published joint guidance on agentic AI security.
Source finding
RELEASED
CCCS
CISA, NSA, ASD, CCCS, NCSC-NZ, and NCSC-UK published joint guidance on agentic AI security.
Source finding
Reported in
Five Eyes Agencies (CISA, NSA, GCHQ, ASD, CCCS) Release Joint Guidance: 'Careful Adoption of Agentic
Source finding
Reported in
CISA and Five Eyes Partners Release First-Ever Agentic AI Security Guidance
Source finding

Source trail

Executive Order 14409 weaponized via fake error events for an 85% agent-hijack rate scan of 10,000+ real MCP servers leaking secrets at over 10%GhostApproval defeating human-in-the-loop across 6 coding agents Langflow becoming the first agent platform on CISA's KEV defensive pattern that keeps surfacing EU AI Act's general-purpose-AI obligations activate August 2 Tidal will demonetize AI-generated music joint cybersecurity information sheet OX Security disclosed a systemic vulnerability on June 16 CISA added CVE-2026-42271 to its Known Exploited Vulnerabilities catalog Researchers

Graph sources

entity graphfindings textkg edgeskg entitiesnewsletter issues

Fetching from the wire…