Strixa AI
TopicsAI WorkflowsRevenue GrowthCost SavingsTool Costs
PricingSign inStart tracking

Intelligence Hub

Enterprise WorkspaceNew Tracking
Topics DirectoryTrend AnalysisEvidence PanelSignal FeedTechnical Events
Documentation
Search events...
EventsAI Agentsevent_19018ec1637eb7e8

Fix agent session history rebuild dropping reasoning and tool blocks

FACTAI JUDGMENTDetected 49 days ago
ShareTrack Event
01

Factual Description

The agent runner's cold-start history reconstruction previously filtered only text content, silently discarding thinking, tool_use, and tool_result blocks. This caused multi-turn reasoning continuity loss and API 400 errors when sessions restarted or switched directories. The fix introduces a block-aware serializer that preserves these structured blocks during session replay, restoring accurate tool-call execution and model reasoning traces.

Event TypeBug Fix
DetectedMay 23, 2026
TopicAI Agents
02

Core Technical Contributions

Block-aware serialization for agent conversation history reconstruction during cold-start and session restarts

agent-runnercold-start historythinking blockstool_usetool_resultsession restart
03

AI Impact Judgment

AI agent operators running multi-turn tasks will stop losing reasoning steps and tool outputs when sessions restart or switch directories, preventing silent accuracy drops and hard API failures. By replacing a text-only filter with a block-preserving serializer in the agent runner, the system now correctly replays structured reasoning and tool tags into the prompt preamble. This stops providers like DeepSeek V4 Flash from rejecting requests with 400 errors and restores explainability for models relying on chain-of-thought context. Watch for potential context window bloat if the new serialization processes large nested tool calls or file attachments, and verify that downstream parsers handle the updated XML structure without breaking existing integrations.

Confidence0%
Importance85
Evidence1
04

Raw Evidence Links

Github Pull RequestOpenCoworkAI/open-cowork PR #224: fix(agent-runner): preserve thinking blocks when rebuilding cold-start history

The previous rebuild loop did: msg.content.filter((c) => c.type === 'text') which silently dropped thinking, tool_use, and tool_result blocks.

Event Contextevent_19018ec1637eb7e8
ID
event_19018ec1637eb7e8
Entity Map
agent-runner / cold-start history / thinking blocks
Confidence Score
0% Watching
Observer Node
ai_agents
Processing Latency
Batch observed

Maturity vs Risk Vector

MaturityCode
Risk FlagsContext Window Bloat / Parsing Compatibility / Nested Tool Call Edge Cases
Confidence0%

Raw JSON Payload

{
  "event_id": "event_19018ec1637eb7e8",
  "topic_id": "ai_agents",
  "event_type": "Bug Fix",
  "event_time": "2026-05-23T08:15:36Z",
  "title": "Fix agent session history rebuild dropping reasoning and tool blocks",
  "summary": "The agent runner's cold-start history reconstruction previously filtered only text content, silently discarding thinking, tool_use, and tool_result blocks. This caused multi-turn reasoning continuity loss and API 400 errors when sessions restarted or switched directories. The fix introduces a block-aware serializer that preserves these structured blocks during session replay, restoring accurate tool-call execution and model reasoning traces.",
  "contribution": "Block-aware serialization for agent conversation history reconstruction during cold-start and session restarts",
  "impact": "AI agent operators running multi-turn tasks will stop losing reasoning steps and tool outputs when sessions restart or switch directories, preventing silent accuracy drops and hard API failures. By replacing a text-only filter with a block-preserving serializer in the agent runner, the system now correctly replays structured reasoning and tool tags into the prompt preamble. This stops providers like DeepSeek V4 Flash from rejecting requests with 400 errors and restores explainability for models relying on chain-of-thought context. Watch for potential context window bloat if the new serialization processes large nested tool calls or file attachments, and verify that downstream parsers handle the updated XML structure without breaking existing integrations.",
  "maturity": "Code",
  "confidence": 0,
  "importance_score": 0.85,
  "risk_flags": [
    "Context Window Bloat",
    "Parsing Compatibility",
    "Nested Tool Call Edge Cases"
  ],
  "evidence_count": 1
}

Internal Feedback

Sign in to submit review notes for this event judgment and its evidence trail.

Strixa AI
TopicsAI WorkflowsRevenue GrowthCost SavingsTool Costs
PricingSign inStart tracking