Fix reply_text[:200] truncation breaking bench keyword matching #3

New Issue

alvis · 2026-03-24T01:56:50Z

alvis commented

2026-03-24 01:56:50 +00:00

Problem

agent.py now logs reply_text: {final_text[:200]} — truncated to 200 chars. Bench recall keyword checks in test_memory.py parse reply_text from this log line, so long replies with keywords past char 200 silently fail.

Fix

Either remove the truncation in agent.py or update the bench to read reply text from a different source.

## Problem `agent.py` now logs `reply_text: {final_text[:200]}` — truncated to 200 chars. Bench recall keyword checks in `test_memory.py` parse `reply_text` from this log line, so long replies with keywords past char 200 silently fail. ## Fix Either remove the truncation in `agent.py` or update the bench to read reply text from a different source.

alvis referenced this issue

2026-03-24 02:42:00 +00:00

Fix tier logging: capture actual_tier, fix parse_run_block regex, remove reply_text truncation #11

alvis referenced this issue from a commit

2026-03-24 02:42:02 +00:00

Fix tier logging: capture actual_tier, fix parse_run_block regex, remove reply_text truncation

alvis referenced this issue

2026-03-24 02:47:53 +00:00

Fix tier logging: capture actual_tier, fix parse_run_block regex, remove reply_text truncation #15

alvis closed this issue

2026-03-24 02:52:44 +00:00

Sign in to join this conversation.