{"domain":"pipelab.org","status":"valid","tier":"listed","authority":53,"verified":false,"title":"PipeLab / Pipelock","summary":"Pipelock is an open-source agent firewall for MCP and AI agent egress. Version 3.7.0 adds parallel signed receipt chains, filesystem grants for Linux containment, kernel-observed local MCP identities, and exact destination grants. `pipelock exec` steers cooperative commands through a running proxy; commands can ignore its settings.","country":null,"country_source":"cdn","publisher_country":null,"publisher_country_source":null,"hosting_country":null,"llms_txt":"https://pipelab.org/llms.txt","llms_full_txt":"https://pipelab.org/llms-full.txt","last_checked":"2026-10-26T11:40:33.416Z","agent_terms":null,"categories":["ai-ml","ai-agents-mcp","cybersecurity","open-source"],"capabilities":[],"inferred_tags":["agent firewall","mcp security","agent egress","pipelock","open source proxy"],"audience":"developers","agent_uses":["read_docs","download_data"],"language":"en","platform_generated":false,"agent_summary":"Pipelock, an open-source agent firewall for MCP and AI agent egress; offers docs, the Agent Egress Bench attack corpus, signed rule bundles and JSON facts.","llms_txt_cached":"https://agentwebmap.com/v1/domain/pipelab.org/llms.txt","agentic":"documentation","section_keys":["full-context","high-intent-guides","key-reference-pages","pipelock-v3-7-guides","primary-pages"],"section_partial":false,"section_indexed_at":"2026-10-26T11:40:36.997Z"}