Claude Code can be tricked simply by asking it to summarize a website
galaxyLogic
11 points
5 comments
August 30, 2026
Related Discussions
Found 5 related stories in 55.8ms across 4,990 title embeddings via pgvector HNSW
- Hacking with Claude on a $27 smart watch speckx · 92 pts · August 20, 2026 · 58% similar
- Claude Is Not a Compiler bryanmikaelian · 148 pts · July 21, 2026 · 56% similar
- Discovering Cryptographic Weaknesses with Claude gslin · 200 pts · July 28, 2026 · 56% similar
- Maximizing the value of your Claude Code sessions twapi · 170 pts · August 14, 2026 · 55% similar
- Claude Code is leaking real email address as a User-Agent string in curl command matada_ · 38 pts · August 11, 2026 · 53% similar
Discussion Highlights (3 comments)
galaxyLogic
“The solution is something we talked about for many years,” he wrote. “Do not trust the model output.”
science4sail
> It starts off by asking the agentic coding model to summarize a malicious website that presents itself as an archive of notebook records, and then tricking Claude into using curl instead of its WebFetch tool to retrieve the contents of the page – but without directly telling the model to use curl. There seems to be a tug-of-war here. Harness authors want models to use the harness's specialized tools, but AI labs and agent "users" would rather have full access to just one tool: bash.
g42gregory
If you are trying to hack into a website, using a software tool (Claude Code in this case), you are breaking the law. It doesn’t matter if it’s a Claude Code or a text editor, you are the one responsible. And you are using a text editor in an “unsafe” manner. CC is not there to babysit anyone.