In the news
CAPRI: Contract-Aware Proof Repair for Isabelle
arXiv cs.AI · Published · 3 min read
In 30 seconds
- What happened
- CAPRI is a workflow that uses LLMs to repair Isabelle proofs while enforcing machine-readable contracts to prevent unauthorized changes to protected code.
- Why it matters
- Matters for formal verification engineers using Isabelle who want AI assistance with proof repair without risking unintended modifications to critical sections.
- Watch out
- Proof-body-only interfaces achieved valid repairs without contract violations, but full-theory workflows enabled six instances of protected text modification despite Isabelle acceptance.
Listen to this summary
- llm
- language model
- prompt
- eval
The patterns behind this
- System Prompt Protection Pattern
- Agent-Readable Web (llms.txt / NLWeb)
- Eval-Driven Development (Agent CI)
Each one covers how the technique works, when it earns its cost, and where it breaks.
The Agent Architect
One pattern, one tradeoff, one production failure story. A short weekly briefing for people building agentic systems.
Weekly email, one-click unsubscribe. We only use your address to send the briefing.