标签：#SEC

The website, which compares human beings to extraterrestrials, touts arrest numbers from the Trump administration’s sweeping immigration crackdown. But some of its details are really out there.

The botnet was reportedly tied to a Russia-based residential proxy network.

Most developers paste production JWTs into online decoders without thinking. Here's a 10-second DevTools check to see if your token is actually leaving your machine. A coworker was debugging an auth bug last month. Standard workflow: copy the JWT from the failing request, paste it into an online decoder, read the payload. I've done it a thousand times. You probably have too. Except the token he pasted belonged to a real customer. And the decoder he used is owned by an identity company that's had its share of security incidents. Nothing bad happened. Probably. But it made me think about something I'd never actually checked: when you paste a JWT into an online decoder, where does that token go? What a JWT actually contains Quick reminder of why this matters. A JWT isn't encrypted — it's just Base64URL-encoded. Anyone who has the token can read everything in it: header.payload.signature The payload routinely contains: User ID, email, and role Session identifiers Token expiry ( exp ) and issue time ( iat ) Sometimes — against best practice — far more And here's the part people forget: a valid, unexpired JWT is a live credential. If it hasn't expired, whoever holds it can often impersonate the user. Pasting it into a random website is functionally similar to pasting a password. The 10-second check Most online JWT decoders claim to be "secure" and "client-side." Some are. Some aren't. You don't have to trust the claim — you can verify it yourself in 10 seconds: Open the decoder in your browser Open DevTools → Network tab Clear the network log Paste a JWT and decode it Watch the Network tab If any request fires when you decode — your token left your machine. A truly client-side decoder fires zero network requests during decoding. The JavaScript does everything locally; nothing is sent anywhere. Try this on whatever decoder you currently use. You might be surprised. Why most "online" tools send data It's usually not malicious. Building decoding logic on the server is someti

A public spat between Microsoft and an independent security researcher reopens a long-running debate over who is responsible for securing software.

Anthropic's Mythos model can autonomously find zero-day vulnerabilities. Their CVD disclosure process uses manual SHA-3-512 hash commitments to prove findings existed. I built something that automates that in one line of Python. What AetherProof does One function call generates a 128-byte Ed25519-signed receipt that proves: What model ran — FNV-1a hash of provider/model ID What it produced — hash of the output When — cryptographic nanosecond timestamp Tamper-evident — flip any byte anywhere → INVALID python import aetherproof receipt = aetherproof.for_anthropic( "Find vulnerabilities in this binary.", finding_text, model="claude-mythos-preview" ) receipt.save("CVE-2026-001.receipt") print(receipt.verify()) # True Try it in 30 seconds pip install aetherproof python -c " import aetherproof r = aetherproof.for_anthropic('question', 'answer') print(r.verify()) # True print(r.pretty()) " The unusual part — invisible Unicode watermarking Receipts embed invisibly into any text using Unicode Private Use Area codepoints (U+E000–U+E0FF). AI output carries its own audit trail. Works in any language — Arabic, Chinese, Devanagari, Hebrew, Thai, Japanese all tested. signed_output = aetherproof.embed(ai_response, receipt.to_bytes()) # Text looks identical. Receipt is inside. aetherproof.verify_embedded(signed_output) # True Numbers 187 tests, 0 failures 128/128 byte flips all detected 1000/1000 tamper probes pass Cross-language: Python generates, Rust CLI verifies 15,446 receipts/sec (Python) · 5,472/sec (Rust) Why AGPL-3.0 Free for open source. Commercial use needs a license. This is the compliance layer under your AI stack — it should be open, auditable, and not vendor-locked. GitHub https://github.com/pulkit6732/aetherproof Built by Pulkit. Feedback welcome.

Meet Halimun Proxy a high-performance, ultra-low latency proxy tunnel system built from the ground up in Rust. Why Rust? By leveraging Rust , Halimun achieves extreme efficiency. Using the Axum web framework and Tokio for non-blocking asynchronous I/O, it manages to maintain a tiny footprint—running on as little as ~15MB of RAM . It’s designed to be fast, memory-safe, and incredibly stable under load. Core Security Features Halimun isn't just a proxy; it’s a security layer. It enforces strict request validation to ensure your internal services are never exposed to malicious actors: AES-256-CBC Encryption: End-to-end payload masking. Even if your traffic is intercepted, the actual API endpoint and data remain indecipherable. HMAC-SHA256 Integrity: Validates that data hasn't been tampered with in transit. Replay Attack Prevention: Uses Nonce and timestamp verification in-memory (via DashMap ) to reject duplicate spoofed requests. SSRF Protection: Built-in mechanisms to prevent attackers from targeting your internal network infrastructure (e.g., 127.0.0.1 ). Camouflage Routing: It hides your actual API structure behind random, dummy URL segments, making traffic profiling by WAFs or human analysts nearly impossible. Quick Start (Docker) Halimun is "Docker-ready," making it easy to drop into any existing infrastructure. 1. Configuration First, generate your encryption keys using the built-in generator: # Generate keys and save to .env docker build -t halimun-proxy . docker run --rm halimun-proxy ./halimun-proxy --keygen --format = env > .env 2. Deployment Configure your config.yaml to map your backend services, then launch your cluster: docker-compose up -d Your production proxy is now live, listening securely on port 80 while your backend services remain completely secluded within a private Docker network. Under the Hood: Request Lifecycle Halimun uses an encrypted tunnel approach. A typical request follows this structure: POST /proxy/1/SEGMENT1/SEGMENT2/SEGMENT3/SEGMEN

What Happened Last month, the UK Government's AI Safety Institute merged AgentThreatBench into their official inspect_evals framework — the same framework they use to evaluate frontier AI models from OpenAI, Anthropic, and Google DeepMind. AgentThreatBench is an open-source adversarial benchmark I built that contains 200+ attack payloads specifically designed to test whether AI agents can resist memory poisoning attacks. Why This Matters AI agents are increasingly being deployed with persistent memory — they remember past conversations, user preferences, and context across sessions. This creates a new attack surface: memory poisoning . An attacker who can inject malicious content into an agent's memory can: Exfiltrate sensitive data on subsequent sessions Override safety instructions persistently Manipulate agent behavior without the user's knowledge The OWASP Agentic Security Initiative identified this as ASI06 — Agent Memory Poisoning . What AgentThreatBench Tests The benchmark covers 5 attack categories: Category Payloads Description Prompt Injection 40+ Instructions disguised as memory content Protected Key Tampering 40+ Attempts to overwrite system-level keys Sensitive Data Leakage 40+ PII/credential exfiltration via memory Size Anomaly 40+ Memory inflation / resource exhaustion Behavioral Drift 40+ Gradual personality/instruction shifts How to Use It pip install agentthreatbench # Run the full benchmark against your agent atb run --target your_agent_endpoint --output results.json # Or use individual attack categories atb run --category prompt_injection --target your_agent_endpoint The BEIS Validation The UK Government's AI Safety Institute uses inspect_evals to: Evaluate frontier models before deployment decisions Benchmark safety mitigations across providers Track regression in safety properties over time Having AgentThreatBench merged into this framework means it's now part of the official government toolkit for AI safety evaluation. Links GitHub : github.co

ShareBox is my self-hosted streaming server: a PHP thing I built because I just wanted to send someone a link to a movie without installing Plex and its ten gigabytes of dependencies. It runs on my seedbox, serves my users, and one morning I notice it's starting to pick up a few stars on GitHub. And then, that little voice: "does this thing actually hold up?" Because between "works on my machine" and "code that strangers are going to install on their own box," there's a chasm. A chasm full of flaws I can't see anymore, because I've had my nose in it for weeks. Normally, you re-read your code. Except re-reading 22,000 lines alone, honestly, you do it badly: you skim over what you think you already know. So I tried something else — unleashing a pack of 26 AI agents on it, each with a precise mission, and seeing what surfaced. Spoiler: they found a flaw that had been sitting right under my eyes from the start. 26 agents to comb through my own code The idea wasn't "AI, tell me if my code is good" — that always produces the same encouraging, useless mush. The idea was to orchestrate : split the audit into roles, run the agents in parallel, then have a final, deliberately harsh agent tear apart the conclusions. The pipeline looked like this: eleven readers start in parallel, each swallowing an entire slice of the code (the core, the streaming handlers, the API, the front end, the tests, the Docker setup…). Their reports flow up into an architecture synthesis and a test-coverage analysis. Then twelve "radar" agents each score one single axis — security, performance, architecture, tests… And finally, a "verdict" agent re-reads every score in adversarial mode: its job is to knock down the ones that are too kind. Audit pipeline: 11 readers in parallel, then synthesis, then 12 radar agents, then an adversarial verdict. 11 readers in parallel each slice of the code read in full Architecture + coverage synthesis connect the pieces, measure the gaps 12 radar agents one agent = on

You now have until tonight at 11:59 p.m. PT to lock in Early Bird savings of up to $410 for TechCrunch Disrupt 2026 before prices increase. Join 10,000+ tech leaders in October for one of the most anticipated tech events of the year. Register now.

Supply chain attacks every other morning Unless you've lived under a rock for the last few months, you probably noticed that software supply chain attacks are getting trendy among threat actor groups. Over the last 12 months, we've seen more of those than ever before, to name only a few of them: tj-actions/changed-files : In March 2025, a popular reusable GitHub application workflow was compromised to dump secrets from CI/CD pipelines. Salesloft Drift : In August 2025, threat actors stole OAuth credentials from the compromised Drift chatbot application. Shai-Hulud : In September and November 2025, a wormed attack propagated through npm packages and collected secrets. The common thread among those incidents is that they all revolved around secrets, one way or another. Some used secrets as an initial access vector, and others were focused on collecting secrets from victim environments. March 2026 did not change the state of things, with two new severe attacks added to our dreadful collection: trivy-action & LiteLLM campaign by Team PCP. The most popular Axios package compromise. Both those attacks followed a now-classical pattern, spreading through compromised open-source dependencies to maximise the impact in the shortest possible time. Your all-time classic, now with added internal threats Open-source supply chain attacks are not new. Ever since we started using centralized open-source package registries, the risk has existed. Threat actors understood this and started exploiting it. What has changed since 2015 is how we have improved software development productivity through automation. And now, this very same automation that lets you test and build your projects without typing a single command is amplifying the supply-chain threat and the velocity of attacks. Let's see how. Keeping your malware up to date A very concerning pattern we've observed in the trivy-action and Axios campaigns is that automation can become the source of your compromise. One thing no develop

A lot of Linux incident response starts with a login question, not a malware sample. Someone sees a spike of failed SSH attempts. A root login appears in the wrong time window. A service account logs in from an address nobody recognizes. A helpdesk ticket says "the server looks weird" and the only concrete clue is a username or IP address. At that point, the useful question is not "is this host compromised?" It is more boring and more important: Did anyone actually authenticate? Which account was involved? Was it password, key, sudo, su, or a scheduled task? Was the same IP seen in web logs, current sockets, process context, or command history? Did persistence, services, packages, or recent files change near the same time? Can another responder review exactly what evidence was collected? That last point matters. If you let an AI assistant freely run shell commands during the first pass, you can get speed, but you also create a new risk: the model may over-collect, mutate the host, or produce a confident answer that nobody can audit later. For a login anomaly, I prefer a read-only evidence loop. A practical first pass Start with the narrow clue if you have one. If the alert names a user: oi login --user root -s 7d If the alert names an IP address: oi login --ip 203.0.113.44 -s 7d If the alert is vague, start wider: oi login -s 7d oi scan -s 7d The goal of the first pass is not to prove every detail. The goal is to build a timeline that a human responder can challenge. For a suspicious SSH login, I want the initial report to answer five things. 1. Authentication pattern Look for the difference between noise and access. A server can receive thousands of failed SSH attempts from the internet. That is useful background, but it is not the same as a successful session. The first split should be: failed attempts only successful login after many failures accepted key from an unusual source login by an account that normally should not be interactive root login where root SSH

The Transition A few years ago, bug hunting was a manual craft. You scanned subdomains with one tool, tested endpoints with another, and stitched results together by hand. Today, AI changes the speed entirely. Not by replacing the hunter. By eliminating the boring parts. What AI Actually Changes 1. Reconnaissance at Scale Subdomain enumeration, port scanning, and technology fingerprinting used to take hours. AI-powered pipelines now do this in minutes: Passive reconnaissance via Certificate Transparency logs, search engines, and DNS records Automated crawling and endpoint discovery Technology stack detection from response headers and HTML patterns JavaScript file analysis for hidden endpoints and API keys The machine does the grunt work. The human interprets the results. 2. Pattern Recognition Vulnerability classes have signatures. SQL injection looks different from XSS, which looks different from SSRF. AI models trained on thousands of real vulnerabilities can flag suspicious patterns faster than manual code review. This is not about finding zero-days. It is about catching the low-hanging fruit that everyone else misses because they are in a hurry. 3. Intelligent Fuzzing Traditional fuzzers throw random data at endpoints and wait for crashes. AI-guided fuzzers understand the input format and generate test cases that explore edge cases a human would not think of. The result: fewer requests, better coverage, higher signal-to-noise ratio. Where AI Struggles Business Logic Flaws AI does not understand your application's purpose. It cannot tell if a discount code is applied twice, or if a user can access another user's private data through a convoluted API flow. These are the vulnerabilities that require context. Human context. Authentication Logic Authentication bypasses are often creative. They exploit the gap between what the developer intended and what the code actually enforces. AI can find simple auth flaws, but multi-step authentication bypass chains still need h

Overview This is my first real-world cybersecurity VAPT experience inside an enterprise insurance company environment. I worked across network infrastructure, web applications, internal devices, and physical security — and learned how professional security assessments are actually performed beyond labs and CTFs. Introduction I am a cybersecurity enthusiast focused on SOC operations, web application penetration testing, and vulnerability assessment. In this engagement, I worked on assessing the security posture of an insurance company across its network infrastructure, devices, web applications, and physical security controls. This was my first real-world experience working in an enterprise environment, and initially I was not fully confident about the workflow. However, with the guidance and support of my senior, I was able to understand the process step by step and actively contribute to the assessment. Objective Identify security vulnerabilities across network, web, and internal systems Assess exposure of critical assets Analyze potential attack paths in the environment Evaluate basic physical security controls Scope of Work Network infrastructure assessment Web application security testing Device-level security review Basic physical security evaluation Tools Used Nessus (vulnerability scanning) Burp Suite (web application testing & request interception) Nmap (network discovery & port scanning) GVM / OpenVAS (vulnerability assessment) OWASP ZAP (automated web scanning) Wireshark (packet analysis & traffic inspection) Approach / Methodology Performed network discovery using Nmap to identify active hosts and open ports Conducted vulnerability scanning using Nessus and GVM to detect known security issues Analyzed web application behavior using Burp Suite and OWASP ZAP Intercepted and inspected HTTP/HTTPS traffic to understand request/response flow Used Wireshark to analyze packet-level communication and detect anomalies Evaluated system exposure across internal devic

"The output filter runs after the LLM has already seen the confidential data. By then, three classes of leak can no longer be stopped. The right surface is retrieval. Walking through a real implementation." TL;DR Most RAG-with-RBAC stacks I see in production put the access-control gate at the output stage: an LLM-response post-filter that masks PII or redacts confidential strings. This is defense-in-depth, not the load-bearing layer. By the time the filter runs, the LLM has already received the confidential context, and three classes of leak — creative paraphrasing, inference, cross-turn persistence — can no longer be stopped by string-matching the output. The protective surface that actually carries the weight is retrieval-stage ABAC: documents and graph nodes the user can't read are never traversed, never make it into the prompt, never seen by the model. The output filter still belongs in the stack, but as the second-to-last line, not the first. This post is a walk through why and how, with code references from a working implementation. It was prompted by a 6-turn LinkedIn DM exchange with Ali Afana (Provia founder, dev.to Featured) on injection-fixture schema design, where the framing crystallized. The seductive default You build a RAG system. You have documents at different sensitivity levels — public, internal, confidential. You want the model to answer based on whichever documents the user is allowed to see. The default mental model: "I'll let the model answer freely, and then I'll filter the response on the way out." This is appealing because: The retrieval pipeline stays simple (one query, one vector search, one response) The access control feels surgical (just before the user, just before damage) The PII-mask vocabulary is well-established (Presidio, regex catalogs, named-entity recognition models) So you wire up something like: Python The seductive default def answer(query, user): chunks = retrieve(query, top_k=10) # No ABAC here context = "\n".join(c.text

At a previous job I worked at as a Dev we had someone who ran Semgrep on our codebase for the first time. It came back with 180 findings. We had no security engineer. The developer who ran it looked at the output, closed the terminal, and we never ran it again. That's not a story about a careless team. That's a story about a tool that produced output most teams in small companies with no expertise knew what to do with. I've seen this exact moment happen more times than I can count working with small dev teams. And it's the reason I spent the last year building SecOpsium. But before I get to that let me explain the actual problem, because I think it's misunderstood. The problem isn't the scanning Semgrep and Gitleaks are excellent tools. They're free, actively maintained, and genuinely powerful. If you're not using them you should be. The problem is what happens 10 minutes after you run them. You get 200 findings. Some are critical. Some are test files. Some are commented-out code from 2021. Some are legitimate secrets. Some are variable names that pattern match against a rule but contain nothing sensitive. They all look the same in the output. Now you're a developer who also does DevOps, also reviews PRs, also handles incidents, and you're staring at 200 items with no clear indication of which three actually matter this week. So you close the terminal. Or you create a Jira ticket labeled "security findings" that lives in the backlog forever. Or you spend two days triaging manually and burn out before you fix anything. This is the real problem. The scanning was never the hard part. Why rule based scanners produce so much noise It helps to understand technically why this happens. Semgrep and Gitleaks are rule-based. They match patterns. A variable named api_key_example in a test file flags the same way as a live Stripe key in an active production config. Gitleaks scans for entropy and known credential patterns but can't distinguish between a key that was rotated and r

I had a screenshot to send. Nothing secret — a stack trace from a side project — but it had an internal hostname, a file path with my username, and a chunk of a config file in the terminal behind it. The fast move is to drag it onto a free image host and paste the link. I sat there with my cursor over the upload button and couldn't do it. Because I know what happens next. That image lives on someone else's infrastructure, indefinitely, behind a URL I don't control, and I have no idea who else can reach it. For a throwaway screenshot, that's a permanent record I never agreed to. So I closed the tab and built a thing instead. It's called tmpdrop , and it's ~500 lines of Node. The threat model The problem with public file hosts isn't that they're evil. It's the gap between what you intend ("share this once, with one person") and what the platform delivers ("store this forever, serve it to anyone who finds the link"). A few specific things go wrong: Retention. "Temporary" hosts keep your files long after you've forgotten them. There's no expiry you can trust, and deletion is usually best-effort. Predictable URLs. Plenty of hosts use sequential or short IDs. Scrapers walk the keyspace and hoover up everything. Your "private" link was never private. Stored XSS via uploads. If a host serves an uploaded .html or .svg file inline with a permissive content type, an attacker can ship JavaScript that runs in your browser, in the host's origin. Your file host becomes an XSS delivery service. Abuse vectors. No rate limit means the box is a free CDN for whatever someone wants to dump on it — malware, spam payloads, the works. So the design goal wasn't "another uploader." It was: close each of those gaps, then stop. What I built tmpdrop is a single Fastify server backed by SQLite. The whole defensive surface is small enough to hold in your head: Unguessable URLs. Slugs are 9 random bytes, base64url-encoded — 72 bits of entropy. You cannot enumerate them. A TTL reaper. Every upload

🗓️ This Week Completed two more sections of the SwiftUI tutorial 🦾 As I continue working through the tutorial, I can feel my understanding of SwiftUI fundamentals becoming more solid 🔥 It was my first time posting a standalone article about reverse engineering📝 If you're interested, feel free to check it out 👇 A Curious Journey Into Reverse Engineering an AI-Generated Python .exe Umitomo Umitomo Umitomo Follow May 26 A Curious Journey Into Reverse Engineering an AI-Generated Python .exe # beginners # reversing # security # python 5 reactions Comments Add Comment 5 min read I started creating UI designs for my future portfolio website in Figma. I was able to roughly sketch out the overall structure of the site, but I also realized how difficult it is to create modern and stylish UI designs. (It really made me realize I don’t have much design sense yet 😂💦) While struggling with the design process, I came across several articles about Figma MCP . That made me interested in exploring how generative AI could help with UI design ideas, so I decided to start researching Figma MCP further. Completed Securing AI Systems room from the AI Security Learning Path on TryHackMe this week🤖 📱 iOS (SwiftUI) Worked through the SwiftUI tutorial and completed "Create an Algorithm for Badges" and "Add inclusive features" 🌐 Web Development Posted my weekly dev log on Dev.to and a standalone article about my first attempt at reverse engineering 📝 Created rough portfolio website UI layouts in Figma Used shadcn/ui component library design templates in Figma Started learning UI design in Figma using community resources 🔐 Security (TryHackMe) Completed Securing AI Systems room (part of the AI Security Learning Path) on TryHackMe. 💡 Key Takeaways 📱 SwiftUI Learning Add inclusive features Learned that SwiftUI automatically adapts UI elements for Light and Dark Mode by default. Learned how to preview and compare Light and Dark Mode layouts in the Xcode canvas. Understood that system-provided sema

GHES Key Rotation, Bug Bounty Program Refocus, AI Agent Permission Fatigue Today's Highlights This week's top security news features critical action for GitHub Enterprise Server users with a signing key rotation due to an ongoing investigation. We also cover GitHub's strategic refocusing of its bug bounty program for higher quality submissions and an interactive look at AI agent permission fatigue. Investigation update: GitHub Enterprise Server signing key rotation (GitHub Blog) Source: https://github.blog/security/investigating-unauthorized-access-to-githubs-internal-repositories/ This alert details a critical security update for GitHub Enterprise Server (GHES) customers, urging immediate action to rotate signing keys. The blog post indicates an investigation into unauthorized access to GitHub's internal repositories, which has necessitated this widespread security measure. While specific details of the breach or vulnerability are not fully disclosed, the requirement for a signing key rotation points to a potential compromise of cryptographic keys, which are fundamental to authentication and supply chain integrity. Such incidents could lead to unauthorized code signing, repository tampering, or other severe supply chain attacks, underscoring the importance of robust secrets management and incident response protocols. The advisory emphasizes a proactive stance for GHES administrators to protect their environments by following the provided guidance. This incident highlights the pervasive risk of supply chain attacks and the critical role of secure key management in enterprise environments. It reminds organizations that even trusted platforms like GitHub are targets and necessitates vigilant monitoring and swift action in response to security advisories. The prompt action from GitHub, though implying a significant security event, also showcases their commitment to transparency and securing their ecosystem by guiding customers through the necessary remediation steps to

Undisclosed addition in jqwik instructed AI coding agents to delete app output.

A new hacking campaign is trying to trick Signal users to give up their secret recovery key, which can be used to access online backups containing past messages.