Assess Guardrail Defense not yet reviewed
pipelock
Pipelock is an open-source AI agent firewall that sits between AI agents and the network, inspecting mediated HTTP, WebSocket, MCP, A2A and optionally TLS-intercepted traffic for secret exfiltration, prompt injection, SSRF, tool poisoning and risky tool-call chains. It also emits signed action receipts so reviewers can verify its decisions.
Promising or important to understand; evaluate, don't depend on it yet.
Facts
- Repository
- pipelab-org/pipelock
- Homepage
- github.com
- Listed via
- GitHub topic
- First seen
- 2026-11-08
- License
- Apache-2.0
- Stars
- 987
- Last commit
- 2026-11-08
- Checked
- 2026-11-08
Listing is not endorsement. Summaries and notes are AI-assisted and reviewed; repository facts refresh automatically.