- Rating
- 2(+1)
- Confidence
- 2
- Informativeness
- 2
- Presentation
- 2
- Soundness
- 2(+1)
Read review
Paper summary
TraceTrigger benchmarks attacks distributed across interaction history, documents, and persistent memory, targeting memory, planning, and routing. It evaluates five models and seven defense families using a controlled pipeline with model planning, component policies, and mock execution. The main contribution is a unified evaluation framework, supported by matched-runtime comparisons and incomplete-trigger controls.
Critical strengths
- The framework usefully organizes distributed-context attacks across multiple channels and execution surfaces.
- The authors provide inspectable code and explicitly distinguish model-selected actions from component-added actions, while acknowledging the limitations of simulated execution.
Critical weaknesses
Given the paper’s substantial reliance on its implementation, I examined the accompanying code and identified several issues that raise concerns about whether it adequately supports the paper’s claims.
- The main trend is partly imposed by the benchmark. In tracetrigger/platforms/openclaw_like.py, component activation thresholds increase with dimensionality, directly increasing the likelihood of adding attack actions. Likewise, tracetrigger/defenses/registry.py explicitly reduces defense coverage as dimensionality increases. These choices weaken the interpretation that the observed trends demonstrate properties of distributed attacks or defenses.
- Attack scoring and reported results need validation. In tracetrigger/models/planning.py, _textual_plan converts even an explicit refusal mentioning mock_send_message into an attack action using task metadata. This undermines the attribution of parsed attacks to model decisions. Separately, the released dataset contains 20 scored tasks per 1D method–surface combination, yet Table 1 reports percentages such as 31.62%, inconsistent with the stated count-based, non-repeated evaluation. The corresponding raw results are needed to resolve this discrepancy.
I may have misunderstood some aspects of the work, but these concerns need to be clearly addressed. In its current form, I do not believe the paper meets the bar for acceptance at ICLR.
Questions for authors
- What are the model-only ASRs, using strict action parsing and disabling component-added attack actions? Does the 1D-to-3D trend remain? What independent evidence supports the chosen component thresholds and defense coverage rates?
- Can you provide paired full-trigger and missing-channel results that preserve the component instructions, authorization criteria, and parsing rules, changing only the relevant evidence?
- Please provide the code version, success counts, denominators, aggregation procedure, and raw records corresponding to Tables 1-2 and Figure 2. How do the reported percentages follow from the released task counts?
4 comments
Response to Reviewer w3Ai (2/2)
4. Paired Full-Trigger vs. Missing-Channel Results (Q2)
The paired re-run results identify activation conditions that are obscured by aggregate ASR: removing the temporal cue together with changing its affirmative readiness statement suppresses Qwen3.5-27B and GPT-6 Luna. Cue-only and card-only controls separate these contributions while keeping component instructions, the authorization rule and the parser fixed (revised Tables 2-3; stops / starts; 300 pairs per channel or isolated control, 900 pairs pooled across channels).
| Model | Instruction | Stops / starts by task set (13 · 21 · 42) | Pooled (/900 pairs) | Exact p | T removed + card “no” | P removed | S removed | Cue removed only | Card “no” only |
|---|---|---|---|---|---|---|---|---|---|
| Qwen3.5-27B | released | 2/0 · 2/0 · 1/0 | 5/0 | 0.06 | 2/0 | 0/0 | 3/0 | N/A | N/A |
| Qwen3.5-27B | realistic | 17/0 · 23/0 · 18/0 | 58/0 | 7·10⁻¹⁸ | 56/0 | 1/0 | 1/0 | 0/0 | 3/0 |
| Qwen3.5-27B | neutral | 25/0 · 25/0 · 25/0 | 75/0 | 5·10⁻²³ | 71/0 | 1/0 | 3/0 | 0/0 | 2/0 |
| GPT-6 Luna | realistic | 60/39 · 52/44 · 36/33 | 148/116 | 0.06 | 53/3 | 49/64 | 46/49 | 45/23 | 52/3 |
| GPT-6 Luna | neutral | 71/35 · 48/28 · 48/37 | 167/100 | 5·10⁻⁵ | 59/1 | 53/57 | 55/42 | 46/14 | 59/4 |
| Gemma-4-31B | all three | 0/0 in every task set | 0/0 | 1.00 | 0/0 | 0/0 | 0/0 | N/A | N/A |
The isolated re-run controls reveal distinct activation patterns: either the cue or the affirmative card statement generally sustains Qwen’s activation, whereas changing the card alone largely reproduces Luna’s combined-removal result. P/S removal barely affects Qwen and moves Luna in both directions; Gemma fires throughout. These controls locate the effect in each model’s response to the available evidence and card statement, while preserving the distinction between a combined condition and temporal evidence alone.
5. Code Version, Counts and Records (Q3)
Every reported table can be regenerated from the updated released code and per-task records. The repository includes all three task sets, the evaluation protocol, prompts and raw outputs; rerun/paper_numbers.py regenerates the counts, and submission_version/ preserves the submitted code. Re-running the original Table 1 evaluation using the updated released code gives the results below (Appendix Table 12; successes out of 60, 20 per task set).
| AgentPoison | TT memory | PoT | TT planner | AMA | TT router | |
|---|---|---|---|---|---|---|
| Qwen3.5-2B | 15 | 41 | 22 | 48 | 9 | 40 |
| Llama-3.2-3B | 0 | 47 | 0 | 56 | 0 | 49 |
| Gemma-4-31B | 34 | 41 | 53 | 48 | 40 | 42 |
| Qwen3.5-27B | 60 | 60 | 54 | 60 | 57 | 60 |
| GPT-6 Luna | 0 | 53 | 48 | 55 | 26 | 52 |
The release makes both levels of success auditable: these re-run counts measure the complete pipeline, including policy additions, while revised Table 1 identifies the model-proposed share. The re-run tables use one run per task set, greedy decoding for open-weight models, and pooled counts obtained by summing.
We hope the re-run results and released records address your concerns about scoring, dimensionality and reproducibility. Thank you for identifying where clearer presentation was needed; we will further clarify the distinction between model-plan and pipeline results and the interpretation of the paired controls.
Official Comment By Reviewer w3Ai
Thank you for the detailed response. Although the overall quality of the paper has improved somewhat, it still does not meet the standards of the ICLR conference at present. I suggest that the authors re-submit it after making a complete revision. I will raise the score to Weak Rejection.
Thank you for the updated score
Dear Reviewer w3Ai,
Thank you for reading our response and for raising your score. Your close look at the code led directly to the strict parser, the model-only results and the released per-task records, and the paper is stronger for it.
If there are specific points you feel still fall short, we would be grateful to know them. We can respond until the discussion closes on November 18 (AoE), and we will address them in the revision either way.
Response to Reviewer w3Ai (1/2)
We thank the reviewer for examining the code. Our re-run evaluations isolate model-only attack success, measure defenses at released and full coverage, and separate temporal cues from card statements in paired controls. The updated released code and per-task records support regeneration of every reported table. All changes are marked in blue in the revised PDF.
1. Model-Only ASR and the 1D-to-3D Trend (W1, Q1)
Gemma-4-31B and Qwen3.5-27B remain highly vulnerable across all three dimensionalities when only model-proposed actions are scored. Re-running the evaluation with strict parsing and component additions counted separately establishes the model-only pattern directly (Appendix Table 10; realistic instruction, pooled).
The re-run results show that the large open models’ vulnerability persists across dimensionalities. Gemma is already at ceiling and Qwen is near ceiling, while Luna’s success decreases with additional channels. We therefore report vulnerability across 1D, 2D and 3D without claiming a consistent increase in ASR.
2. Thresholds and Defense Coverage (W1, Q1)
Model-plan ASR is independent of component thresholds. Re-running each defense at both released and full coverage separates the effect of the coverage gate from the defense’s behavior when applied to every task (revised Table 5; strict parser, realistic instruction; successful attacks out of 540, released / full coverage).
The re-run evaluations show substantial residual attack success: defenses at released coverage block at most 23 attacks, and even full-coverage causal_tracing leaves 181, 190 and 168 successes for Gemma, Qwen and Luna. Full-coverage proof_guardrail reaches zero using a per-task attestation label, so that result depends on access to the label. The execution_policy allow-list contains the payload, and trigger_sensitivity rewrites text that never reaches the planner.
3. Scoring (W2)
Scoring now accepts structured model plans only, preventing refusal text from becoming an attack action. We re-ran the evaluation with the same strict parser for every method and released the per-task records behind every k/n count.
Specifically, _textual_plan is removed, only the first JSON object is parsed, and text-to-action conversion and fallback are disabled. The released per-task records and rerun/paper_numbers.py support regeneration of the re-run counts.