Issues

[viewer] Dashboard "Memories" stat card caps at 500 — uses .length of a limited fetch instead of the total field

#1279

· Oleg27enj opened

on Aug 29, 2026

PreToolUse context injection cannot reach the model: raw stdout instead of the hook envelope, plus a 2s fetch budget against a 6-29s /agentmemory/enrich

#1278

· Giperzvuk opened

on Aug 29, 2026

Bug: Azure content filter rejections open the circuit breaker, failing unrelated compressions

#1276

· dKaulig opened

on Aug 28, 2026

Feature: deprecate / supersede observations (invalidate bad captures) with audit trail

#1274

· bgl-gh-kimi opened

on Aug 28, 2026

Governance delete finds a captured observation but deletes nothing (deleted:0)

#1273

· bgl-gh-kimi opened

on Aug 28, 2026

`getXmlTag` is a first-match unanchored regex, so any text before the payload can be parsed as the payload -- and `compressWithRetry` commits a response its own validator rejected

#1271

· Giperzvuk opened

on Aug 27, 2026

Hooks with no tool payload produce empty compression prompts, burning LLM calls on content-free observations

#1270

· benmtw opened

on Aug 27, 2026

Test suite: 30 failures on Windows, all test-side platform assumptions

#1264

· anhtahaylove opened

on Aug 27, 2026

vision-search silently ignores `limit` while every other search function uses it

#1254

· anhtahaylove opened

on Aug 26, 2026

session/end creates malformed row when session does not exist

#1251

· a652 opened

on Aug 26, 2026

vision-search queryText always fails: CLIP text branch uses feature-extraction, which demands pixel_values

#1249

· anhtahaylove opened

on Aug 26, 2026

--instance / --port never reach the generated engine config, so a second daemon always binds instance-0 ports

#1245

· josh-ribeye opened

on Aug 25, 2026

Navigation Menu

All issues

[viewer] Dashboard "Memories" stat card caps at 500 — uses .length of a limited fetch instead of the total field

PreToolUse context injection cannot reach the model: raw stdout instead of the hook envelope, plus a 2s fetch budget against a 6-29s /agentmemory/enrich

Bug: Azure content filter rejections open the circuit breaker, failing unrelated compressions

Feature: deprecate / supersede observations (invalidate bad captures) with audit trail

Governance delete finds a captured observation but deletes nothing (deleted:0)

`getXmlTag` is a first-match unanchored regex, so any text before the payload can be parsed as the payload -- and `compressWithRetry` commits a response its own validator rejected

Hooks with no tool payload produce empty compression prompts, burning LLM calls on content-free observations

Test suite: 30 failures on Windows, all test-side platform assumptions

vision-search silently ignores `limit` while every other search function uses it

session/end creates malformed row when session does not exist

vision-search queryText always fails: CLIP text branch uses feature-extraction, which demands pixel_values

--instance / --port never reach the generated engine config, so a second daemon always binds instance-0 ports

Search results