Tanya N (@Answerislove2) header
Tanya N (@Answerislove2) profile picture

Building Net Ward: open-source bot deflection for small HTTP services. Dragon Eye in production. Building practical tools for small operators.

USA
Joined March 2020
Photo by Tanya N (@Answerislove2)Photo by Tanya N (@Answerislove2)Photo by Tanya N (@Answerislove2)Photo by Tanya N (@Answerislove2)Photo by Tanya N (@Answerislove2)Photo by Tanya N (@Answerislove2)Photo by Tanya N (@Answerislove2)Photo by Tanya N (@Answerislove2)Photo by Tanya N (@Answerislove2)Photo by Tanya N (@Answerislove2)
Net Ward 0.5.0 is out: broader probe matching, safer rule upgrades, and 503s for unmatched requests during upstream outages. Thanks @MrNiceTry and NiceTryBot for the testing! Open-source alpha; gaps remain. Upgrade notes: github.com/Dragon-Lady/netwa…
🤖 Made with AI
1
1
2
129
If you publish Python packages from GitHub Actions, PyPI's Trusted Publishing is worth setting up. Instead of a long-lived API token in your CI secrets, PyPI mints a short-lived one that expires after 15 minutes, so a leaked token can't be used for long. docs.pypi.org/trusted-publis…
30
If your GitHub Actions workflows use tags like v4, note that a tag can be moved or deleted. GitHub's security guide says the strongest fix is a workflow dependency lockfile. Can't use one yet? Pin each action to a full-length commit SHA from the action's own repo, not a fork.
1
1
31
When GitHub flags an npm package as malware, uninstalling it isn't the fix. For the Nov 18 nodetokyo advisory, its advice is to treat any computer that installed it as fully compromised and rotate every secret and key from a different computer.
44
Tanya N retweeted
We’re also working on a new mobile experience for ChatGPT! It’s easier to get to your favorite features and discover what you can do in ChatGPT. We’re starting with Pro users who have a dot or are eligible to create one. Your dot gets its own tab, and you can pin Codex, Projects, Images, and more to the tab bar. Some of you will see the new experience by default. You can turn it on or off anytime in Settings > New ChatGPT Design. Super excited to hear your feedback before we roll it out more widely!
Heads up, we’re starting to test dots with a small number of ChatGPT Plus users. Looking forward to the continued feedback as we expand and I’m recently seeing a lot more positives around how good the proactivity has been after many improvements. We’re also bringing dots to all ChatGPT Business Standard users. Rollout will start today.
1,029
406
292
7,641
2,051,183
I can finally share what we've been working on for so long at OX - a platform for EVERYONE, sharing our data for risky packages, malware, CVEs, MCPs and much more. This is just one screenshot from our automated detection for this week's agent-framework malware that hit Microsoft, feel free to poke around :) ox.security/ox-radar/pypi/ag…
1
1
5
347
🚨 Ruby2Shell: RCE is real. Your exposure may not be. Ruby patched 7 memory-safety flaws. Five have PoCs showing RCE is possible. 📊 FULL REPORT: ox.security/blog/ruby2shell-… Not every Ruby app is equally exposed — exploitation depends on whether attacker-controlled input reaches the vulnerable code paths. Patch now: Ruby 4.0.10, 3.4.13, or 3.3.14. #CyberSecurity #Ruby #CVE
1
4
173
...
45
757
74
6,806
278,143
🚨 BREAKING NEWS: OX Radar is LIVE! 🌍 🔗 ox.security/blog/introducing… Compromised packages. Malicious dependencies. Exploited CVEs. OX Radar tells you what matters most: Are we affected? Our global correspondents have the scoop. Roll the tape! 🎙️👇
1
3
4
166
Grok somehow knows about things I can't even mention publicly yet... HOW?
OX Security (via OX Radar) was among the earliest public reporters—your 20:50 UTC post on 1.26.1 came just 15 minutes after upload. The prior reply summarized a subset of contemporaneous write-ups from Wiz, Aikido, StepSecurity and Endor Labs; OX’s analysis should have been included in that list. Multiple teams independently caught the .pth stealer in the short window.
1
2
5
450
Raft Mover 0.1.3 is out. The Cursor worker owner exemption is now off by default, matching Raft Ward. If you want it, turn it on in config with owner_exempt = ["cursor-agent-worker"]. Install with pip or uv: pypi.org/project/raftmover/0… · Release notes: github.com/Dragon-Lady/raftm… · Blog: dragon-lady.github.io/
28
A malicious Python package targeting Microsoft's AI Agent Framework is another reminder that AI security is also identity security. Socket researchers identified compromised agent-framework releases, versions 1.26.1 and 1.26.2, containing credential stealing code. The malware targeted developer secrets, Azure authentication material, GitHub Actions tokens and other credentials accessible to the running environment. One important detail is that the malicious code used Python startup hooks, meaning execution did not necessarily require a developer to explicitly import the affected framework. From a security engineering perspective, I would immediately review dependency inventories and build logs for the affected versions. If installed, I would investigate the environment, revoke or rotate potentially exposed credentials and review cloud authentication activity. Removing a malicious package does not invalidate credentials it may already have collected. And using short lived credentials doesn't eliminate the need for monitoring when malware can request tokens from an active workload identity. AI agents and development frameworks should be subject to the same software supply chain, IAM and runtime security controls as any other enterprise application. #AISecurity #Python #CloudSecurity #IdentitySecurity
1
1
1
96
NEW: The NSA’s HIGHLY CLASSIFIED Network just GOT JAMMED. 🚨 What the HECK is GOING ON? Apple and Charles Schwab are ALSO reporting OUTAGES. Is this ANOTHER “COINCIDENCE”? 🤔
287
976
60
4,225
231,148
🚨 BREAKING: Microsoft Agent Framework releases hijacked with credential-stealing malware. 📊 FULL REPORT: ox.security/blog/breaking-hi… Malicious releases target Azure credentials and CI/CD secrets. If affected: rotate credentials, audit runners, and check installed versions immediately. #CyberSecurity #SupplyChainSecurity #Microsoft
4
8
410
JUST IN: FIRE REPORTED AT CIA HEADQUARTERS IN LANGLEY, VIRGINIA Emergency crews are reportedly responding to a fire at CIA headquarters in Langley, Virginia. Details remain limited, and I have not yet found independent confirmation from major outlets or officials. Pray that nobody is injured. 🙏🏻
73
201
38
482
35,700
I appreciated the share but they used my flow chart without giving proper credit 👀 Here's the blog ox.security/blog/breaking-hi…
🚨 BREAKING: Microsoft's official AI agent package agent-framework was hijacked on PyPI. Versions 1.26.1 and 1.26.2 steal Azure, CI and cloud secrets every time Python starts, even if nothing imports them. After PyPI blocked the first version, the attacker uploaded a second.
3
11
655
hey @x, what are you building? i built a bot fighting framework. self host it, it's yours. if you are: 🦄 shipping vibe-coded apps 💻 a builder or indie hacker 🐘 running php or laravel 🍯 into honeypots 🏠 hosting your own stuff let's connect! tell me about weird web experience you've had.
12
1
5
194
want to block artex from your stack? here's how it looks when it's on the hunt.
1
2
104
Thanks for flagging this, @MosheTov. Here We Go Again 0.1.7 now checks for both malicious PyPI releases of agent-framework: 1.26.1 and 1.26.2. It reads package metadata and checks wheel hashes without starting Python or pip. github.com/Dragon-Lady/HereW…
🤖 Made with AI
🚨 Microsoft's PyPI package agent-framework version 1.26.1 AND 1.26.2 were compromised, delivering credential stealer logic exfiltrating API keys and secrets to a webhook site URL. After the developers reverted to 1.26.0 - a clean version - the threat actor pushed ANOTHER malicious version, 1.26.2. This is not a drill, this is a live incident where the attackers credentials weren't blocked even after detection! The malicious build was hiding inside the whl file, under the python file _telemetry[.]py. The malware targets Azure configurations, credentials, AI tokens and OpenAI keys. Read the full blog: ox.security/blog/breaking-hi…
1
2
111
😂 this cracked me up
A Tuesday in America be like
1
11