rony-llm-agent/pkg/tools/sandbox
Victor Vargas 07d1840e7e feat(sandbox): network egress policy, secret redaction, untrusted-content fencing
NetworkPolicy validates scheme/host and re-validates resolved IPs at dial
time and on redirects (DNS-rebinding defense), with cloud metadata
endpoints always blocked. Redact masks known credential shapes (OpenAI/
Anthropic/GitHub/AWS/Slack/Google keys, PEM blocks, JWTs) in tool output.
WrapUntrusted fences fetched web content against prompt injection, paired
with UntrustedContentInstruction for the system prompt.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-15 14:50:45 -07:00
..
advanced_test.go feat(sandbox): network egress policy, secret redaction, untrusted-content fencing 2026-07-15 14:50:45 -07:00
network.go feat(sandbox): network egress policy, secret redaction, untrusted-content fencing 2026-07-15 14:50:45 -07:00
redact.go feat(sandbox): network egress policy, secret redaction, untrusted-content fencing 2026-07-15 14:50:45 -07:00
sandbox.go feat(tools): add Tool definitions, Registry pattern, and filesystem sandbox with path validation 2026-06-30 23:53:32 -07:00
sandbox_test.go Fix tool call tracking and streaming assembly for all providers 2026-07-08 16:11:57 -07:00
untrusted.go feat(sandbox): network egress policy, secret redaction, untrusted-content fencing 2026-07-15 14:50:45 -07:00