Agents won’t run or can’t be fixed because “violate our Usage Policy”
Users are reporting that Anthropic's Claude AI is now preventing its agents from running or being fixed, citing violations of the usage policy. One user experienced this when their headless Claude Code agent, which had been functioning for months, suddenly stopped responding. When attempting to have the agent fix itself, Claude refused, indicating a potential policy change by Anthropic regarding such autonomous agent setups. AI
IMPACT Potential restrictions on autonomous agent usage could impact how developers integrate and deploy AI models for complex tasks.