Guardrail decision engine: warning vs hard-stop #670

Closed
opened 2026-10-27 13:37:16 +00:00 by sam · 0 comments
Owner

Implement the guardrail decision engine that uses ToolCallGuardrailConfig thresholds to decide between sending a warning to the LLM or issuing a hard-stop (aborting the tool loop). Considers exact_failure_warn_after, same_tool_failure_warn_after, and no_progress_warn_after. Must track per-session and per-turn counters.


Imported from beads archie-core-1786637493566-145-06408456 · GitHub #166 · type task · priority P2 · created 2026-08-08 · closed 2026-08-08

Implement the guardrail decision engine that uses ToolCallGuardrailConfig thresholds to decide between sending a warning to the LLM or issuing a hard-stop (aborting the tool loop). Considers exact_failure_warn_after, same_tool_failure_warn_after, and no_progress_warn_after. Must track per-session and per-turn counters. --- _Imported from beads `archie-core-1786637493566-145-06408456` · GitHub #166 · type task · priority P2 · created 2026-08-08 · closed 2026-08-08_
sam added the mediumtask labels 2026-10-27 13:37:16 +00:00
sam self-assigned this 2026-10-27 13:37:16 +00:00
sam closed this issue 2026-10-27 13:37:16 +00:00
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: sam/archie#670