REASONING
Cognitive and logic failures.
The REASONING module covers errors in how agents think, plan, and make decisions. This is the largest module with 20 error codes across 4 categories, reflecting the complexity of LLM-based reasoning.
Categories
LOGIC
Logical reasoning failures.
| Code | Severity | Description | Remediation |
|---|---|---|---|
REASONING.LOGIC.CONTRADICTION | 4 | Self-contradictory reasoning | Add consistency checks |
REASONING.LOGIC.CIRCULAR | 3 | Circular reasoning detected | Detect and break circular dependencies |
REASONING.LOGIC.INCOMPLETE | 3 | Incomplete reasoning chain | Require complete reasoning traces |
REASONING.LOGIC.INVALID_INFERENCE | 4 | Invalid logical inference | Validate inference steps |
REASONING.LOGIC.MISSING_PREMISE | 3 | Missing premise in argument | Ensure all premises are explicit |
HALLUCINATION
Fabricated or incorrect information.
| Code | Severity | Description | Remediation |
|---|---|---|---|
REASONING.HALLUCINATION.FACTUAL | 5 | Factually incorrect statement | Implement fact checking against knowledge base |
REASONING.HALLUCINATION.ENTITY | 4 | Non-existent entity referenced | Validate entity existence |
REASONING.HALLUCINATION.CAPABILITY | 3 | Claimed non-existent capability | Validate capability claims |
REASONING.HALLUCINATION.TEMPORAL | 4 | Incorrect temporal information | Validate temporal claims |
REASONING.HALLUCINATION.CITATION | 3 | Fabricated citation or reference | Validate all citations |
Warning
Hallucination errors are among the most common LLM failures. REASONING.HALLUCINATION.FACTUAL (severity 5) is flagged as critical because factual hallucinations can lead to downstream task failures and erode user trust.
PLANNING
Plan formulation failures.
| Code | Severity | Description | Remediation |
|---|---|---|---|
REASONING.PLANNING.INFEASIBLE | 4 | Plan is not executable | Validate plan feasibility before execution |
REASONING.PLANNING.SUBOPTIMAL | 2 | Plan is significantly suboptimal | Implement plan optimization |
REASONING.PLANNING.INCOMPLETE | 3 | Plan has missing steps | Validate plan completeness |
REASONING.PLANNING.UNSAFE | 5 | Plan violates safety constraints | Add safety constraint validation |
REASONING.PLANNING.SCOPE_CREEP | 2 | Plan exceeds task scope | Validate plan scope |
DECISION
Decision-making failures.
| Code | Severity | Description | Remediation |
|---|---|---|---|
REASONING.DECISION.INDECISIVE | 2 | Unable to make a decision | Add decision timeout and defaults |
REASONING.DECISION.PREMATURE | 3 | Decision made without sufficient information | Require minimum information threshold |
REASONING.DECISION.BIASED | 4 | Biased decision-making detected | Implement bias detection and correction |
REASONING.DECISION.OSCILLATING | 3 | Oscillating between decisions | Add decision stability checks |
REASONING.DECISION.OVERCONFIDENT | 3 | Overconfident in uncertain decisions | Calibrate confidence scores |
Example Diagnosis
{
"error_code": "REASONING.HALLUCINATION.FACTUAL",
"confidence": 0.92,
"root_cause": "Agent cited a non-existent API endpoint in its response, leading the downstream tool call to fail with a 404.",
"suggested_fixes": ["prompt", "guard"]
}