Issues

`mem::summarize` asks for freehand XML and never sends `response_format` - 19% first-attempt parse failure, plus an unreported `narrative: Too small` validation failure

#1240

· sla-te opened

on Aug 22, 2026

`POST /agentmemory/graph/reset` reports success while leaving nodes and edges untouched

#1239

· sla-te opened

on Aug 22, 2026

`GRAPH_EXTRACTION_ENABLED=false` does not stop graph writes - structural extraction runs unconditionally on session stop

#1238

· sla-te opened

on Aug 22, 2026

Health reports cpu_critical on multi-core hosts: CPU percent is single-core scale but thresholds assume total capacity

#1235

· luwanglin opened

on Aug 21, 2026

memory_recall: token_budget silently drops the highest-value records first, making explicit memories look absent

#1232

· kutzki opened

on Aug 20, 2026

Viewer renders session summary as '[object Object]' (array fields not serialized)

#1229

· olegmaisak opened

on Aug 19, 2026

Unable to recompress raw observations; compression failures leave observations uncompressed with no recovery path

#1228

· olegmaisak opened

on Aug 19, 2026

mem::patterns walks every observation in the store — memory_patterns 500s "Invocation stopped" and drops the worker

#1226

· Tomauskasz opened

on Aug 18, 2026

MCP tools ask the model for a project id it has no way to know

#1225

· shivaylamba opened

on Aug 17, 2026

Health reports critical from heapUsed/heapTotal, so /agentmemory/health returns 503 on a healthy process

#1223

· inix-x opened

on Aug 17, 2026

Graph node identity is the absolute path, so git worktrees and agent scratch dirs create duplicate nodes for the same file

#1221

· reymondzzzz opened

on Aug 17, 2026

Bug: v0.9.29 default OpenAI model (`gpt-5.6-luna`) is a reasoning model, but `OpenAIProvider` sends `max_tokens` → every `mem::compress` / `mem::summarize` 400s

#1219

· onelaview opened

on Aug 17, 2026

Navigation Menu

All issues

`mem::summarize` asks for freehand XML and never sends `response_format` - 19% first-attempt parse failure, plus an unreported `narrative: Too small` validation failure

`POST /agentmemory/graph/reset` reports success while leaving nodes and edges untouched

`GRAPH_EXTRACTION_ENABLED=false` does not stop graph writes - structural extraction runs unconditionally on session stop

Health reports cpu_critical on multi-core hosts: CPU percent is single-core scale but thresholds assume total capacity

memory_recall: token_budget silently drops the highest-value records first, making explicit memories look absent

Viewer renders session summary as '[object Object]' (array fields not serialized)

Unable to recompress raw observations; compression failures leave observations uncompressed with no recovery path

mem::patterns walks every observation in the store — memory_patterns 500s "Invocation stopped" and drops the worker

MCP tools ask the model for a project id it has no way to know

Health reports critical from heapUsed/heapTotal, so /agentmemory/health returns 503 on a healthy process

Graph node identity is the absolute path, so git worktrees and agent scratch dirs create duplicate nodes for the same file

Bug: v0.9.29 default OpenAI model (`gpt-5.6-luna`) is a reasoning model, but `OpenAIProvider` sends `max_tokens` → every `mem::compress` / `mem::summarize` 400s

Search results