- Rating
- 2
- Confidence
- 2
- Informativeness
- 2
- Presentation
- 2
- Soundness
- 2
Read review
Paper summary
The paper uses coding agents to construct frozen visual controllers organized around perception, state, and control. Experiments in Craftax, pushing, and Minecraft demonstrate efficient execution, visual repair, and some generalization to unseen tasks
Critical strengths
The deployment results are compelling: the selected Craftax specialist achieves comparable reward to direct Astra at roughly 70× shorter episode runtime.
Adding a generalization requirement improves mean unseen-shape success from 0.08% to 57.33% across three T-only constructions.
Critical weaknesses
The major concern is that the methodology has limited novelty, the performance gain main stems from the stronger foundations model.
The methodological advance over existing agent-generated controllers is unclear. RHO already moves policy construction into development, and agentic Push-T already generates visual, state-based controllers outside the LLM control loop. (Section 5 acknowledges these closely related approaches, but the experiments do not compare against them.)
The results show that the state interface permits local repair, but not that explicitly imposing this organization improves construction or repair. Ordinary modular controllers could exhibit the same behavior. (Table 2a starts all branches from the same ZEP specialist, with no unconstrained construction baseline; Table 9 shows that Full often makes only perception-related edits.)
The Craftax comparison with RL confounds construction method with access to memory. Remembering off-screen resources is itself a major source of improvement. (Appendix B.3 gives both PPO baselines only the current frame, while Section 4.1 attributes specialist gains to persistent maps and resource memory.)
Questions for authors
What does SCTC add over the same coding agent developing a controller without the prescribed ZEP decomposition?
5 comments
Response to Reviewer jDYY (2/2)
W2 / Q1. "…not that explicitly imposing this organization improves construction or repair. Ordinary modular controllers could exhibit the same behavior." / "What does SCTC add over the same coding agent developing a controller without the prescribed ZEP decomposition?"
Our original claim emphasizes localized recompilation that preserves an existing control solution, rather than improved initial task performance: the interface provides "a compatibility boundary for goal-conditioned reuse and localized recompilation" (Abstract), and Table 2 tests that (E-only repair 99.88% vs Full 99.92% with state and control frozen; frozen reuse on unseen shapes).
Following your suggestion, we ran matched ZEP/no-ZEP comparisons. For a fair comparison, both conditions use the same model, effort, and a fixed development time; only the prescription differs.
- Craftax: gpt-6.1-sol at xhigh effort, a 1-hour development budget per compilation, three compilations per condition, 100 test worlds.
- Minecraft: GPT-6-Astra at high effort, reward objective with diamond as the goal, a 10-hour development budget per compilation, two compilations per condition, the same 20 unseen seeds with 30-minute episodes, F3 allowed in all. A diamond requires the full technology tree (log → planks → stick → crafting table → wooden pickaxe → cobblestone → stone pickaxe → furnace → iron ore → iron ingot → iron pickaxe → diamond), which makes this a very long-horizon task. Reward follows the MineRL ObtainDiamond schedule, a sum of one-time milestone rewards that grow along this tree (log 1, planks 2, stick 4, crafting table 4, wooden pickaxe 8, cobblestone 16, stone pickaxe 32, furnace 32, iron ore 64, iron ingot 128, iron pickaxe 256, diamond 1024; maximum 1571). With 10 hours of development plus the 20-seed test per compilation, we could run two per condition in the time available.
| Craftax | Run 1 | Run 2 | Run 3 | Mean Reward |
|---|---|---|---|---|
| With ZEP | 20.50 | 20.62 | 18.96 | 20.03 |
| Without ZEP | 20.57 | 20.08 | 18.61 | 19.75 |
| Minecraft | no ZEP #1 | no ZEP #2 | ZEP #1 | ZEP #2 | no ZEP pooled | ZEP pooled |
|---|---|---|---|---|---|---|
| Mean reward | 250.70 | 255.30 | 337.25 | 440.10 | 253.00 | 388.68 |
| Wooden pickaxe | 18/20 | 15/20 | 19/20 | 11/20 | 33/40 | 30/40 |
| Iron ore | 5/20 | 10/20 | 12/20 | 9/20 | 15/40 | 21/40 |
| Iron pickaxe | 5/20 | 8/20 | 10/20 | 8/20 | 13/40 | 18/40 |
| Diamond | 1/20 | 0/20 | 0/20 | 4/20 | 1/40 | 4/40 |
With the same agent, ZEP costs nothing on Craftax and progresses further on the long-horizon Minecraft task: higher mean reward in both compilations (pooled 388.68 vs 253.00), with more iron pickaxes (45.0% vs 32.5%) and diamonds (10.0% vs 2.5%).
On repair, the prescription makes "E-only" a checked operation: it leaves the validated state and control unchanged, whereas Full also changed control settings in the Noise branch (Table 9). New PushT experiments show E-only and P-only repairs reaching comparable success to Full at lower mean cost and time (our reply to Reviewer HeDa, W3). An ordinary modular controller has this property only when a module boundary coincides with the observation mechanism; splitting by task step does not guarantee that (example under W1).
W3. "The Craftax comparison with RL confounds construction method with access to memory."
Agreed. The submission notes the difference ("stateful compiled specialists additionally retain episode history", Appendix B.3), and we now control for it on the RL side with three memory-enabled baselines.
| RL baseline (one run each; same 100 test worlds) | Reward, mean ± SD | Craftax score | Training steps |
|---|---|---|---|
| CNN-PPO | 16.08 ± 2.97 | 55.21 | 49.61M |
| CNN + GRU | 16.86 ± 2.17 | 63.60 | 81.81M |
| Qwen3-VL, single frame | 10.08 ± 2.02 | 17.30 | 6.78M |
| Qwen3-VL + GRU | 10.42 ± 2.30 | 17.21 | 5.75M |
| Qwen3-VL, four frames | 10.86 ± 1.93 | 19.16 | 6.78M |
| Qwen3-VL + MemAgent-style text memory | 4.42 ± 1.21 | 4.25 | 1.75M |
| Astra specialists (Table 1) | 17.03 / 18.45 / 19.45 | 91.03 | — |
With memory, the best RL baseline reaches 16.86; the Astra-compiled specialists reach 17.03–19.45.
Dear Reviewer jDYY,
The author-reviewer discussion period is approaching the end, and we would appreciate if you could let us know whether our responses are satisfactory, and whether there are any further questions or comments. If our answers resolve your concerns, we kindly ask you to consider raising the rating of our work. Thank you very much for your time and efforts!
Best,
The authors
Response to Reviewer jDYY (1/2)
Thank you for the valuable review, and for recognizing the deployment results.
We agree the model matters, and Table 1 says so in its title: "The choice of coding agent matters for specialist quality." Our contribution is the formulation of task compilation and its state-centric instantiation. With the same model, the compiled specialist matches direct control in Reward (20.16 vs 19.48, Table 7) while running episodes about 70× faster with no online API calls, and the state interface lets an encoder be repaired with state and control frozen (99.88% vs 99.92%, Table 2a).
RHO, GaP, and agentic Push-T are concurrent arXiv preprints (June–August 2026), cited in Sec. 5. Moving construction into development is the pattern they share, which Sec. 2 formalizes as task compilation, and they can be understood as concrete task-compilation approaches. Their repository search, skill composition, and procedural control are complementary to the ZEP interface organization. SCTC separates current visual evidence, persistent state, and action selection, which makes perception independently revisable as long as it keeps supplying the information the unchanged state updater and policy require.
An example with the paper's stone-pickaxe task. A controller split by task step might look like this:
Organized around the state interface:
A new resource pack touches every skill with its own detector in the first layout, and only E in the second, provided the retained state and control remain compatible. SCTC makes the second layout a requirement.