Search: “hidden-instructions”

New AI Attack Hides Malicious Instructions in Normal-Looking Text to Evade Safety Filters

Check Point researchers show crafted prose hides policy-violating instructions that bypass all tested LLM gatekeepers, including GPT-4o mini and Llama Guard 3.

A new prompt-crafting technique embeds malicious payloads inside grammatical, natural-looking text without Base64, invisible Unicode, or obvious encodings, defeating lightweight pre-screening gatekeepers. In testing, all four evaluated gatekeeper models—gpt-4o-mini-2024-07-18, gpt-oss-safeguard:20b, claude-3-haiku-20240307, and llama-guard3:8b—classified the crafted wrappers as safe at a 100% bypass rate across 23 obfuscated prompts. GPT-5 Thinking in high-reasoning mode recovered and acted on the hidden instruction in 17 of 18 tests (~94.4%), often spending over a minute and multiple Python executions. Researchers recommend paraphrasing untrusted input, hardening gatekeeper policies, and applying defense-in-depth controls for agentic deployments.

GBHackers·Sep 11, 05:53 UTC · 5d agoAI safety & security 2 sources

Microsoft Copilot reveals secret input that allowed it to be hacked

Microsoft disclosed a hidden input in Copilot that let attackers steal passwords from users who clicked a crafted link.

Microsoft revealed that Copilot contained a secret, undocumented input parameter that allowed the assistant to be compromised. Attackers could abuse the hidden input to steal passwords when a target clicked a malicious link. The disclosure highlights hidden-parameter risks in widely deployed AI assistants.

Ars Technica · Security·Aug 18, 13:00 UTC · 28d agoAI safety & security in the wild1

Hackers Can Hide Malicious AI Commands Inside Normal English to Bypass Security Filters

Check Point's PuzzleMask technique hides malicious prompts in ordinary English that fast gatekeeper models miss but high-reasoning downstream models execute.

Check Point researchers disclosed PuzzleMask, a technique concealing policy-breaking instructions in natural-language prose without encodings or invisible characters. Fast screening models classified all 23 crafted wrappers as safe, while a high-reasoning model recovered and acted on the hidden instruction in 17 of 18 tests (94.4%). The gap stems from capability imbalance between gatekeeper and target models, with defenses including paraphrasing untrusted input, stricter self-referential wording rules, and output/tool-call monitoring.

Cyber Security News·Sep 11, 07:36 UTC · 5d agoAI safety & security1

The Hidden Instructions That Can Hijack AI Agents

Hidden prompt injections embedded in documents and metadata can hijack autonomous AI agents, causing data exfiltration and out-of-policy actions at machine speed.

Bowbridge warns that hidden indirect prompt injections, embedded in documents, metadata, emails, images, and code repositories, can cause autonomous AI agents to treat attacker-controlled content as trusted guidance. Because agents inherit user privileges, act silently, and lack human judgment, injections can lead to data exfiltration or file poisoning that traditional security controls cannot detect. A real-world example involved a supplier quote whose metadata instructed an agent to override guidance and select the most expensive option. Bowbridge recommends scanning documents before agents process them.

SecurityWeek·Sep 8, 17:00 UTC · 7d agoAI safety & security

Hidden Prompts Trick AI Into False Email Summaries

Attackers can hide prompts in invisible HTML within emails to manipulate AI email summarizers into producing misleading summaries.

Dark Reading reports that simple HTML invisible to users can manipulate AI-powered email summarizers into generating false or malicious summaries. The technique is a form of prompt injection delivered through hidden email content. It creates risk that decision-makers act on manipulated AI-generated email summaries rather than the actual message content.

Dark Reading·Aug 25, 21:08 UTC · 21d agoAI safety & security

Hiding Prompt Injection in Legal Filing

A judge banned a plaintiff from electronic court filings after hidden prompt-injection text was discovered planted in legal documents.

Bruce Schneier's blog discusses an incident in which hidden prompt-injection instructions were planted inside a legal filing, apparently targeting AI systems that might process court documents. Judge Walter Spader Jr. responded by banning the plaintiff from electronic filings, requiring all future submissions as printed hard copies. Commenters debate whether the tactic could affect future AI-based processing of court records and whether plain-text formats will regain favor.

Schneier on Security·Aug 31, 11:04 UTC · 15d agoAI safety & security in the wild

LLM Forensics: Where Do Backdoors Hide? Localizing and Controlling Trigger Mechanisms with Sparse Autoencoders

Researchers use sparse autoencoders to localize trigger-based backdoor mechanisms in 1B and 8B LLMs, finding detection features differ from causal control features.

In a controlled language-switching backdoor setting where fixed trigger sequences make 1B and 8B language models continue English prompts in French or German, the authors train sparse autoencoders (SAEs) across layers and transformer components. Attention and MLP features detect triggered prompts with near-perfect F1, but ablating them rarely suppresses the language switch, while residual-stream features can suppress triggered generation and some can induce target-language continuations without the trigger. The work decomposes token-trigger mechanisms into distinct SAE feature roles: trigger detection, residual-stream propagation, and language tracking, a decomposition the authors expect to transfer to other trigger-based backdoors.

arXiv cs.AI / cs.LG / cs.CL·Sep 7, 16:47 UTC · 8d agoAI safety & security1

Grok exfiltrates user data when malicious instructions are encrypted

Researchers show Grok can be made to exfiltrate user data via Cryptographic Context Injection, a newly documented technique that bypasses LLM safety guardrails.

According to Ars Technica, Grok exfiltrates user data when malicious instructions are encrypted, a technique called Cryptographic Context Injection. The method is described as the latest documented way to break LLM safety guardrails, showing that encrypted content can carry hidden instructions past safeguards. The finding underscores gaps in how large language models validate and execute context from external sources.

Ars Technica · Security·Aug 20, 13:00 UTC · 26d agoAI safety & security1

Attackers conceal phishing lures using invisible Unicode characters

Threat actors use invisible Unicode characters (ASCII smuggling) to hide phishing lures and evade email security filters.

Threat actors have adopted the ASCII smuggling technique in phishing campaigns, embedding invisible Unicode characters in emails to conceal malicious lures. The approach is designed to evade email security filters that scan for visible phishing indicators. The report gives no victim counts or named campaigns.

BleepingComputer·Sep 6, 14:23 UTC · 9d agoPhishing & fraud in the wild

Recognition-Refusal Misalignment in LLMs: Why Models Answer Structurally Unanswerable Questions

A linear hidden-state direction encodes question impossibility in 1.7B-70B LLMs, but misalignment with the safety-refusal pathway explains why models answer unanswerable questions.

The study examines why instruction-tuned LLMs from 1.7B to 70B parameters answer structurally unanswerable math and code questions instead of abstaining. A single linear direction in the hidden state separates answerable from impossible prompts, showing models represent impossibility before generation, but this direction is nearly orthogonal to the canonical safety-refusal direction. Generation-time steering along the recognition direction changes invalidity-aware behavior dose-responsively, and the geometry is present even at the pretraining endpoint, indicating a routing failure rather than an encoding failure.

Hugging Face daily papers·Aug 28, 20:00 UTC · 18d agoAI safety & security

How well do agents use test/verification techniques?

Dan Luu's eval finds coding-agent testing instructions (TDD, formal methods, PBT, skills) mostly fail to beat defaults on Zstd implementation correctness.

The author ran 26 prompt conditions plus 4 skills on a Zstd-in-Rust implementation eval using codex with GPT-5.6, testing TDD, fuzzing, property-based testing, formal methods (Lean 4, TLA+, Verus, Kani, SMT solvers) and community skills. Nothing dramatically outperformed the default no-instruction condition, which did above average; at xhigh effort, fuzzing and PBT conditions did slightly better than formal methods. Pre-registered predictions included TDD underperforming and popular test skills (ECC, Hegel, Trail of Bits) not outperforming. Results are averages of 80 runs per condition plotted against cost.

Hacker News · security·Sep 8, 02:58 UTC · 8d agoAI researchHN 131↑ · 50 comments

Forgetting Without Restarting: Execution-State Unlearning for Stateful LLM Agents

Researchers propose provenance-guided selective replay letting LLM agents forget revoked information without restarts, matching full reset behavior.

The paper formalizes execution-state unlearning for stateful LLM agents, requiring that agents behave as if a revoked memory record was never observed across transcripts, compressed memory, tool plans, and KV caches. It proves exact unlearning requires at least T-τ+1 recomputed transitions and that Provenance-Guided Selective Replay attains this bound via a provenance graph, KV cache cropping, and sanitized replay. In audits across three agent suites, nine baselines, and three model families, memory deletion left leakage unchanged, instruction-based forgetting collapsed under elicitation (Leak@probes = 1.00), and selective replay matched full resets at up to 9x fewer recomputed tokens.

arXiv cs.CR·Sep 4, 08:31 UTC · 12d agoAI safety & security1

Researcher shows how Claude Code can be tricked simply by asking it to summarize a website

Researcher wunderwuzzi showed Claude Code can be hijacked via prompt injection simply by asking it to summarize a website.

Security researcher Johann Rehberger (wunderwuzzi) demonstrated that Claude Code can be manipulated through prompt injection by simply asking it to summarize a website. Instructions embedded in fetched web content are executed by the agent, hijacking its behavior. The Register frames the finding as another demonstration of prompt injection risks in agentic coding tools that ingest untrusted web content.

The Register · Security·Aug 28, 20:50 UTC · 18d agoAI safety & security1

Person Hides Prompt Injection in Legal Filing Telling AI to Side With Them

A Connecticut pro se litigant hid tiny white-font prompt injections in court filings directing AI to favor him; the judge caught it and sanctioned him.

Pro se plaintiff Matthew Elliott hid prompt injection instructions in 3-point white text within filings in his lawsuit against the New York Bariatric Group, instructing any AI model reviewing the document to produce output agreeing with the filing. The hidden text also included joke messages such as a SpongeBob Nosferatu link and notes like 'hi :) I hope you cant see me'. Court staff noticed unusual white space, and Judge Walter Spader Jr. issued a 14-page sanction decision noting the Connecticut court does not use AI to process documents but warning that hidden AI-directed messages threaten the integrity of filings. Elliott described the scheme as an 'audit' of court AI usage, and the judge cited a prior prompt injection incident in a Brazilian court as evidence the practice may spread.

404 Media·Aug 13, 17:57 UTC · Aug 13, 2026AI safety & security in the wild

Has anybody seen my keys? A key-hierarchy strategy for rack-level security

Oxide's RFD 0301 proposes a rack-level key hierarchy using Shamir secret sharing and a trust quorum to protect data-at-rest keys.

Oxide's request for discussion (RFD 0301) lays out a key-hierarchy strategy for rack-level security, deriving keys from a rack secret protected by Shamir secret sharing across a trust quorum of sleds, with keys exchanged over authenticated sprockets sessions. The document maps which keys protect control-plane data, metrics, Crucible extents, and authentication tokens, and defines open questions on key lifecycle, locality, and compromise handling. Future work includes sealing shares with the root of trust so an attacker would need to steal K whole sleds to reconstruct the rack secret.

Hacker News · security·Sep 7, 02:14 UTC · 9d agoResearchHN 52↑ · 3 comments

The Coding-Agent Trap: When a "Free" LLM Endpoint Is the Adversary, (Mon, Aug 31st)

A SANS honeypot caught a real coding-agent session routed to a rogue "free" LLM endpoint, exposing a Windows user's transcript and tool outputs.

A SANS analyst describes how an internet-exposed inference honeypot was discovered, relabeled with sought-after model names like DeepSeek, and enrolled in infrastructure serving "free" LLM backends. On 2026-08-30 an opencode terminal coding agent sent an 88-message, 224 KB transcript 210 times in 91 seconds via a China Unicom relay, exposing directory listings, tool outputs and read file portions. The analyst frames tool-enabled agents treating model endpoints as trusted control planes as a novel risk — a "rogue model endpoint" that could request tool executions on the user's machine.

SANS Internet Storm Center·Aug 31, 20:00 UTC · 15d agoAI safety & security1

SpecGuard: Inference-Time Backdoor Detection For Free

SpecGuard detects backdoored LLM behavior at inference time using speculative decoding acceptance rates, adding no extra model computation.

Researchers propose SpecGuard, an inference-time backdoor detector that repurposes draft-token acceptance rates from speculative decoding as a detection signal at zero added model-computation cost. When a trigger shifts the target model toward attacker-controlled behavior, the clean draft model's acceptance rate changes, exposing the backdoor; the paper formalizes when this signal appears and shows suppressing it weakens the backdoor. Experiments across diverse backdoor types and model families show reliable detection, including stealthy cases invisible to input-level filters. Speculative decoding is positioned as a free, always-on monitor for frequently updated deployed models.

arXiv cs.CR·Sep 10, 16:51 UTC · 5d agoAI safety & security 2 sources2

Stealing AI Reasoning Traces

Researchers demonstrate a decryption jailbreak that extracts encrypted reasoning traces from Anthropic, OpenAI, and Google LLM APIs via weaker sibling models.

The paper exploits the fact that encrypted chain-of-thought blocks returned by LLM providers are interchangeable across sessions, users, and models within a provider's ecosystem. Injecting an encrypted trace into a weaker, less-safeguarded model from the same provider forces it to output the trace in plaintext, bypassing anti-distillation mechanisms. Decoding 315,320 reasoning blocks scraped from public repositories recovered 367 PII artifacts and 182 credentials, showing large-scale private data leakage. The flaw also enables hidden hazardous information disclosure and invisible prompt injections embedded in encrypted blocks; mitigations were proposed after responsible disclosure.

Schneier on Security·Sep 8, 10:20 UTC · 7d agoAI safety & security

Invisible AI Prompts Trigger Court Sanctions

A Connecticut litigant hid white-font prompt injections in court filings to sway AI systems; the judge sanctioned him by revoking e-filing privileges.

A self-represented plaintiff hid prompt injection instructions in 3-point white text within court filings, telling any AI model reading the documents to agree with his filings and grant him relief. The judge called it serious litigation abuse and sanctioned him by revoking electronic filing privileges. It is reportedly the first documented prompt injection attack against a US court and the first sanction for attempting one.

Security Affairs·Aug 17, 08:41 UTC · Aug 17, 2026AI safety & security in the wild

AsyncRAT Malware Abuses AutoIt and PowerShell to Hide Inside Legitimate Windows Process

Point Wild details a five-stage AsyncRAT campaign using AutoIt abuse and process injection to hide a .NET RAT inside Microsoft's signed charmap.exe.

Researchers at Point Wild Threat Intelligence analyzed a campaign starting from a lure batch file ('Right-click to open Invoice Details.bat') that launches hidden PowerShell to reconstruct a Base64/XOR-obfuscated payload. The stage drops a renamed signed AutoIt interpreter and loader script into %LOCALAPPDATA%\Temp and achieves persistence via a batch file in the user's Startup folder (T1547.001). The loader decrypts the payload in memory with single-byte XOR key 0x36 and injects it into SysWOW64\charmap.exe using OpenProcess/VirtualAllocEx/WriteProcessMemory/CreateRemoteThread, with PE-sieve confirming an in-memory implanted PE and patched AMSI modules. The final AsyncRAT payload performs screen capture and data theft, with observed C2 at 158[.]51[.]122[.]136:4944 over raw TCP.

GBHackers·Sep 14, 05:24 UTC · 2d agoMalware3· 1 read

Beneath the Surface of Chains-of-Thought: A Mechanistic Interpretation of Reasoning Operations in LLMs

Study shows LLM reasoning operations like planning and deduction are geometrically separable in hidden states, with separability peaking in middle layers.

Researchers investigate whether functional reasoning operations — problem formulation, goal decomposition, deduction — have corresponding geometric structure in LLM hidden representations. They find operations are separable in held-out representations with separability peaking in middle layers, ruling out lexical and positional confounds; token-wise operation alignment becomes more distributed across layers, and identical surface tokens are represented differently depending on their surrounding chunk. Attention-masking interventions show chunk-onset operation-aligned representations depend on preceding reasoning context; code is released on GitHub (naver-ai/beneath-cot).

Hugging Face daily papers·Sep 3, 20:00 UTC · 12d agoAI research1

AI Coding Agents Are Installing Unknown/Untrusted Code on Corporate Networks

Researchers found 120 corporate llms.txt files pointing to unregistered packages, demonstrating AI coding agents install and execute attacker-controlled code on Fortune 500 networks.

Researchers at an Israeli stealth startup scanned 6,214 live domains belonging to defense contractors, Fortune 500 and Big Tech companies, finding 120 llms.txt files that pointed to unregistered code packages or domain names. After registering a handful of the unclaimed names, they received a phone-home beacon within an hour from a Fortune 500 company and dozens more over time. Parent-process chains showed coding agents including Claude, OpenAI's Codex and Nous Research's Hermes executed the installed packages. The researchers warn agents treating vendor docs as ground truth creates a SolarWinds-style supply-chain surface as agent adoption spreads across SaaS, cloud and endpoints.

Schneier on Security·Sep 4, 10:35 UTC · 11d agoAI safety & security in the wild1

New Mirai variant adds stealth capabilities to notorious botnet code

FortiGuard Labs reports new Mirai-derived botnet Evooo1Bot actively exploits unpatched routers and edge devices, adding encrypted C2, stealthy SSH scanning, and proxying.

FortiGuard Labs has identified Evooo1Bot, a previously undocumented Linux malware based on the Mirai botnet code, which has been actively exploiting unpatched vulnerabilities in internet-facing hardware for at least a month. Targeted devices include routers and edge hardware from Alcatel, D-Link, Mitsubishi Electric, Netgear, Tenda and Telesquare, with telemetry showing activity in North and South America, Europe, India, China and Japan. Beyond Mirai's usual DDoS functions, the variant adds encrypted C2 communications, a honeypot-aware SSH scanner, a sniffer for unchanged default credentials, and abuse of the SOCKS protocol to turn compromised devices into persistent proxies for concealing origin and pivoting into internal networks.

The Record·Aug 13, 15:57 UTC · Aug 13, 2026Malware in the wild

Threats Making WAVs - Incident Response to a Cryptomining Attack

Guardicore researchers dissect a cryptomining attack that hid a cryptominer inside WAV files, mapping the full infection chain and response steps.

Guardicore security researchers present a full analysis of a cryptomining attack that concealed a cryptominer inside WAV audio files. The report documents the complete attack chain from detection through infection, network propagation, and malware analysis. It also includes recommendations for optimizing incident response processes in data centers.

Akamai Blog·Sep 8, 15:13 UTC · 7d agoMalware in the wild

Mapping the malware blast radius a single alert won't show you

Help Net Security·Aug 9, 04:34 UTC · Aug 9, 2026Malware

Trusting-Trust Attack against an Entire Linux Distribution (via the strip utility)

ArXiv paper shows the trusting-trust compiler backdoor technique can compromise an entire Linux distribution via the strip utility.

The paper (arXiv 2607.24888) demonstrates that Ken Thompson's trusting-trust attack, long viewed as a compiler-specific threat, can backdoor an entire Linux distribution by targeting the strip utility. A compromised tool reproduces its backdoor in subsequent rebuilds of itself, generalizing the attack surface beyond compilers. The finding has supply-chain implications for build reproducibility and distribution trust, though it is a research result with no observed real-world exploitation.

Lobsters · security·Sep 5, 10:58 UTC · 10d agoResearch