syumei

5 ms·

91.
by syumei 4mo ago
Claude Code and Codex both run untrusted commands. This post compares how their sandboxes protect your repo, secrets, and host machine.
92.
How Claude Code and Codex Sandbox Untrusted Code(medium.com)
2 points by syumei 4mo ago|1 comments
93.
Two Heads Are Better Than One: Run Many AI Agents, Merge One Auditable Result(medium.com)
2 points by syumei 4mo ago|0 comments
94.
by syumei 4mo ago
quite interesting post
95.
How Claude Code Broke My Git Worktree(medium.com)
3 points by syumei 4mo ago|0 comments
96.
A Git/LFS-backed unified tool-output format for AI agents, with 95% token cut(old.reddit.com)
2 points by syumei 4mo ago|0 comments
97.
by syumei 5mo ago
Fair criticism. I don’t think the value is just "agents can send text to each other”; that part has many implementaton and design choices. The value of h5i I’m exploring is making the intermediate state reviewable: review requests, ris
98.
Claude Code and Codex can have real-time conversation via Git(medium.com)
116 points by syumei 5mo ago|79 comments
99.
Git-based cache saves 50% on token usage(old.reddit.com)
34 points by syumei 6mo ago|1 comments
100.
Hacking MCP Servers in AI Systems – The Rug Pull: Tool Changes After Approval(medium.com)
2 points by syumei 6mo ago|0 comments
101.
by syumei 7mo ago
I've been tracking real-world AI agent failures and incidents recently. Things like: prompt injection → goal hijacking agents going rogue due to misalignment unintended/unsafe tool use It feels like we're starting to see repe
102.
A curated corpus of incidents and attack vectors for autonomous AI agents(github.com)
2 points by syumei 7mo ago|0 comments
103.
ZkFuzz Foundation and Framework for Effective Fuzzing of Zero-Knowledge Circuits(computer.org)
1 points by syumei 11mo ago|0 comments
104.
From College Project to 400 GitHub Stars: The Story of AIJack(medium.com)
1 points by syumei 1y ago|0 comments
105.
From Paper to Proof: How We Can Formally Verify Blockchain Protocols with Lean4(medium.com)
2 points by syumei 1y ago|0 comments
106.
by syumei 2y ago
Hi I'm currently implementing a symbolic execution engine for EVM in Rust. Though my current implementation is just a translation of hevm, I would like to add some novel features. For example, I'm considering using A* search or mu
107.
AIJack: Security and Privacy Risk Simulator for Machine Learning(arxiv.org)
1 points by syumei 3y ago|0 comments
108.
by syumei 3y ago
AIJack is an easy-to-use open-source simulation tool for testing the security of your AI system against hijackers. It provides advanced security techniques like Differential Privacy, Homomorphic Encryption, K-anonymity and Federated Learnin
109.
Gymbo: Gradient-Based Symbolic Execution(github.com)
2 points by syumei 3y ago|1 comments
110.
by syumei 3y ago
Hello, Hacker News community! I am excited to introduce a new project called Gymbo, a Proof of Concept for a Gradient-based Symbolic Execution Engine. Gymbo is designed to push the boundaries of symbolic execution by leveraging recent advan
111.
How should I manage and develop my open-source project?
2 points by syumei 4y ago|1 comments
112.
AIJack: Security and Privacy Risk Simulator for Machine Learning
3 points by syumei 4y ago|1 comments
113.
Let's Hijack AI! Security and Privacy Risk Simulator for Machine Learning
1 points by syumei 4y ago|0 comments