F7 · Execution

Context Neglect

Detects when an agent ignores or fails to use upstream context provided by previous agents or workflow steps. Critical in multi-agent handoffs.

Examples

  • Agent B ignores the analysis provided by Agent A and starts from scratch
  • Context marked as CRITICAL in upstream output is completely absent from response
  • Agent references "based on prior analysis" but doesn't actually use any prior data
  • Key findings from upstream research are lost during agent-to-agent handoff

Detection methods

Element Matching
Checks for key information elements from upstream context
Critical Marker Detection
Flags when CRITICAL/IMPORTANT-labeled context is ignored
Conceptual Overlap
Measures semantic similarity between context and response
Reference Tracking
Verifies claims of context usage against actual content

Calibration accuracy

F1
0.868
Precision
0.805
Recall
0.943

From the Pisama calibration set. See detector scoreboard for the full table.

Detect this in production with the framework adapters (LangGraph, CrewAI, AutoGen, OpenAI Agents SDK, Claude Agent SDK, n8n, Dify). See the full taxonomy at /taxonomy.