Search results

hindsight-openclaw: recall silently skipped on default agent:main:main sessions — "dispatch surface $CHANNEL does not match session provider main"

Status: Open.

#1182 In vectorize-io/hindsight;

· LaFleurAdvertising opened on Apr 20, 2026

Reflect fails to answer when using Qwen models via LM Studio, succeeds with llama-server

bugSomething isn't workingSomething isn't working

Status: Open.

#1179 In vectorize-io/hindsight;

· mwhardesty opened on Apr 20, 2026

create_mental_model: surface tags_match override and warn about all_strict default (follow-up to #786)

Status: Open.

#1171 In vectorize-io/hindsight;

· normalnormie opened on Apr 20, 2026

reflect() + mental model generation ignores reflect_mission identity framing — always outputs third-person documentary voice

Status: Open.

#1159 In vectorize-io/hindsight;

· grimmjoww opened on Apr 20, 2026

MCP update_bank silently drops mission field and doesn't expose disposition / reflect_mission / retain_mission / observations_mission

Status: Open.

#1156 In vectorize-io/hindsight;

· normalnormie opened on Apr 20, 2026

daemon start race condition: N concurrent session starts spawn N daemons (v0.5.3)

Status: Open.

#1155 In vectorize-io/hindsight;

· el-analista opened on Apr 19, 2026

Add optional access-key login for the Hindsight Control Plane

Status: Open.

#1148 In vectorize-io/hindsight;

· eightHundreds opened on Apr 19, 2026

hindsight-memory: SessionEnd hook does not flush un-retained turns before daemon stop

Status: Open.

#1145 In vectorize-io/hindsight;

· cometto2007 opened on Apr 18, 2026

Gemini Batch API support + extend Batch API to consolidation pipeline

Status: Open.

#1144 In vectorize-io/hindsight;

· FitzDegenhub opened on Apr 18, 2026

bug: OpenAI-compatible embeddings batch size is hardcoded to 100 and not configurable

Status: Open.

#1142 In vectorize-io/hindsight;

· eightHundreds opened on Apr 18, 2026

Support gemini-embedding-2-preview via Vertex AI

enhancementNew feature or requestNew feature or request

Status: Open.

#1139 In vectorize-io/hindsight;

· Gentoli opened on Apr 17, 2026

Prompt in fact_extraction.py misleads LLM into using Markdown instead of strict JSON

Status: Open.

#1138 In vectorize-io/hindsight;

· Asrafiel opened on Apr 17, 2026