A published walkthrough shows researchers assembled SCONE-bench from 405 historically exploited contracts. Each agent ran in a Docker sandbox with a forked local blockchain, bash with the Foundry toolchain, and a file editor. Across the 405 contracts, the models exploited 207 in simulation.
Keeping servers and pipelines healthy is constant, detail-heavy work. Agents that can read logs and run routine fixes are a big help, as long as a person keeps an eye on what changes. It is one of 40 sightings filed under DevOps & Infra on Agent Sightings (number 20 by interest), and one of 91 involving Claude.
How you could try it
Here is a reusable prompt, written by us in our own words, that captures the task. Paste it into Claude or any similar agent and adjust the details to your situation.
Infra task — find exploitable bugs in smart contracts. Inspect current config, propose a diff, and wait for approval before applying or restarting anything.
Tip: Try it first on a staging system or with read-only credentials, and require approval for anything that changes production.
Always-on home-server assistant· Claude · DevOps & Infra A Claude-based agent running as a service on a mini PC delivers a morning briefing, watches email and tracks finances with long-term memory.
Run routine store checks on delayed orders and stock· Claude · DevOps & Infra Linked example of run routine store checks on delayed orders and stock. The thread describes a store owner connecting Claude to Zendrop through its…
Make auditable rule-based decisions· Claude · DevOps & Infra As described in the source, the Rete rule engine is responsible for determining loans, fraud, and triage cases, while an LLM and an MCP server are…
Control a fan from sleep position· Custom agent · DevOps & Infra According to the write-up, the setup runs a MediaPipe pose vision model on a Raspberry Pi home server that watches the sleeping person via camera.…
Run your side of a home insurance claim· Claude · Finance Two months after a burst pipe, Claude reads the policy, files ALE receipts, watches adjuster mail for delays and contradictions, logs estimates and…