Model or Harness? A 41-Failure-Mode Taxonomy That Localizes Where Agent Systems Break

Model or Harness? An Interaction-Centric Taxonomy for Localizing Agent Failures

Harsh Raj, Vipul Gupta, Anas Mahmoud, Razvan-Gabriel Dumitru, Darvin Yi, Aakash Sabharwal, Yunzhong He

cs.AI

2026-07-31

A 41-mode taxonomy from Scale AI localizes each agent failure to an edge between two components and a fault side, telling you whether to fix the model, harness, or environment; top judge reaches κ=0.76.

What problem this solves

When an agent run fails, the log shows a system-level outcome: the task didn't finish, the wrong email went out, the code didn't run. But the same visible symptom can call for very different fixes: retraining the model, reworking the harness (tool integration, memory, context management), fixing a buggy evaluation environment, or repairing the benchmark itself. The authors call this the "repair-assignment problem": identical failures, different root causes, different fixes.

Most existing failure taxonomies are fine-grained lists built for a single benchmark and do not transfer across systems. This paper wants a shared structure that works across architectures and is directly actionable: once you classify a failure, it tells you which component to change.

Method

The authors decompose an agent system into a focal model plus surrounding components, grouped into three families:

Each failure is localized to an edge between two components and tagged with a fault side, which end of the edge owns the repair. Model-side failures point to post-training targets; harness-side failures point to tool-integration or context-management fixes; environment- or grader-side failures mean the evaluation conditions themselves must be redesigned.

The attribution rule matters: trace backward from the system-level failure to the "earliest failure from which execution does not recover" and treat that as the root cause; later errors count as consequences. Of the 41 failure modes, 36 are model-attributable and 5 land on surrounding components.

Results

The taxonomy is grounded in 40 worked examples (SWE-bench, ClawsBench, Harbor-Mix, Anthropic system cards, GitHub issues, agent trajectories). A representative case: an agent deleted more than 200 emails, but the root cause was not the model "wanting" to delete them. Context compaction dropped the constraint telling it to leave the inbox alone, a textbook context-side harness fault.

Reproducibility is tested with an agent-as-judge setup: each judge receives a reference to the failure source (via Docent or Face), reconstructs the evidence itself, then classifies. Four frontier models serve as independent judges:

Judgeκ vs human category labels
GPT-5.50.76
Claude Opus 4.60.71
Claude Opus 4.70.71
Claude Opus 4.80.70

Pairwise agreement peaks at κ = 0.84 (Opus 4.6 vs 4.8). Requiring all four judges to agree lifts category precision to 0.96, but coverage drops to 68%.

Why it matters

For anyone building agents, this answers "where to look after a failure," not just "did it fail." The fault-side split maps straight onto which tool or team should intervene, and it transfers across architectures: coding assistants (Claude Code, Codex), long-running personal assistants that read mail and browse, and multi-agent systems. A judge agreement of κ = 0.76 suggests the categories capture genuine shared structure rather than one annotator's preferences.

Limitations

Thirty-six of the 41 modes land on the model, a heavy model-side skew the authors acknowledge; this risks underweighting harness and environment failures, which in practice are where many agent problems actually live. The 40 worked examples are curated, not a random sample of failures, so there is selection bias toward clear, interesting cases. A κ of 0.76 is "substantial" agreement, far from perfect, and category-level agreement is better than fine-grained failure-mode-level. The taxonomy is a localizer, not a fixer: it names the edge, not the specific patch. Finally, the "earliest unrecovered failure" rule gets ambiguous in long trajectories where many errors interleave.

Terms

Source

What people are saying

Related papers

All paper explainers