Proposal Explorer

Add claims to see which ones the mechanisms address.

In production

• R3 Model identity attestation for showing users that a service runs the declared model weights

Built for an adversarial prover

Network taps and certifiers, Bounding unexplained information in outputs and Bandwidth limits and compartmentalization

No new hardware needed

Model identity attestation

Flaws since mitigated

• Launch-state attestation does not by itself cover weights loaded later in Model identity attestation ^{1 13}
• Verifier dictionary attacks on hashes in Network taps and certifiers ¹⁵

Published attempts to break a system, including those that found failures. Testing history does not show that open flaws are resolved.

• Model identity attestation: Independent red-team
• Network taps and certifiers: Analysis
• Bounding unexplained information in outputs: Independent red-team
• Bandwidth limits and compartmentalization: Analysis

Open critical flaws

• Underlying attestation can be forged or relayed in Model identity attestation

  Critical for the enclave route against an operator with physical access to affected hardware, or control of an unpatched SEV-SNP hypervisor. It does not apply to the recomputation route. PAL*M excludes physical attacks, and Tinfoil acknowledges this boundary.

  The enclave route inherits the platform-specific TEE attestation failures. Intel TDX forgery and H100 relay were demonstrated with physical access and host control. Battering RAM defeated AMD SEV-SNP attestation on DDR4 servers; RMPocalypse did so from malicious host software on platforms without AMD's fixes. These demonstrate failures of the trust roots, not of each model-commitment protocol. ^{2 4 8 10 11 12}

  Response: The TEE.fail authors report that physical interposer attacks are outside Intel's and AMD's threat models. AMD reports fixes for RMPocalypse.

  Demonstrated attack · Critical · Open · Inherited finding. On the record · Related finding in TEE remote attestation for AI workloads

  Related mechanism R1 Hardware-enabled guarantees (flexHEG) and guarantee processors: A tamper-protected enclosure around the chip is the proposed answer when the party that holds the hardware may attack it physically. A pointer, not evidence that this flaw is mitigated. Add

Open significant flaws

16 flaws in 4 mechanisms

• For private models, a user can confirm consistency but not content in Model identity attestation

  When weights are not published, users can check that the same root hash is served each time, but not what the model is. Pairing the hash with an attested evaluation, as in Attestable Audits, is one proposed remedy. ^{1 14}

  Open question · Significant · Open. On the record

• Recomputation depends on trusted logging and randomness, and its tolerance leaves a covert channel in Model identity attestation

  The recomputation variant assumes that every input, output and seed is logged correctly, and that the attacker can neither predict nor manipulate which messages are sampled for verification. Legitimate nondeterminism concentrates at a few token positions, and slow leaks within the tolerated slack remain possible. An independent study showed that an adversary who controls the prompts roughly doubles the bits leaked per token, reducing the exfiltration slowdown from 146–254 times under benign prompts to 60–118 times. The attack targets the exfiltration bound, not the check that outputs match the declared model. ^{3 9}

  Demonstrated attack · Significant · Open. On the record

  Related mechanism R1 Network taps and certifiers: Taps are proposed to copy and hash traffic on the monitored links, reducing reliance on the prover's own log. This still depends on the monitored boundary and trusted capture. A pointer, not evidence that this flaw is mitigated. In the proposal.

  Related mechanism R3 Deterministic and bit-exact inference: Bit-exact inference would remove the numerical tolerance that leaves this channel. A pointer, not evidence that this flaw is mitigated. Add

• Output nondeterminism leaves covert capacity in Network taps and certifiers

  Hashing cannot remove information hidden in the outputs themselves. The Secure Gateway Device paper estimates that about 0.1 bit per token remains even with seed-synchronized replay checks. For a 200k-GPU inference cluster at full load (2,000 tokens per GPU per second), that is about 40 Mbit/s of covert egress, enough to move a 1 TB model in under three days. The paper names this the core remaining challenge and points to deterministic replay or active scrubbing of hardware-induced entropy. An independent study found that an adversary who chooses the prompts roughly doubles the bits leaked per token under Gumbel-based inference verification; see Bounding unexplained information in outputs. ^{9 15}

  Theoretical argument · Significant · Open. On the record

  Related mechanism R3 Deterministic and bit-exact inference: Deterministic replay is one of the two remedies the flaw's source names. A pointer, not evidence that this flaw is mitigated. Add

  Related mechanism R2 Bounding unexplained information in outputs: Bounds the hidden information outputs can carry by measuring what the declared computation fails to predict. A pointer, not evidence that this flaw is mitigated. In the proposal.

• Some links cannot be passively tapped in Network taps and certifiers

  Cankaya notes that copper-connected scale-up domains (for example NVL72 racks and TPU v7 cubes) are much harder to tap than fibre, and that optical budgets make passive taps impractical on 400GBASE-SR8 multimode links. Amodo found no taps advertised for 53 GBaud links as of May 2026. ^{16 23}

  Open question · Significant · Open. On the record

• Encrypted fabrics hide plaintext from both parties in Network taps and certifiers

  Cankaya notes that with TEE-protected sessions whose keys are ephemeral and managed inside the TEE, neither the operator nor the manufacturer can recover session keys after the session, so tapped traffic could not be opened for recomputation. For other encrypted fabrics, the operator can retain keys. ¹⁶

  Open question · Significant · Open. On the record

• Residual side channels in simple passive setups in Network taps and certifiers

  Amodo's analysis of its own tapped prototype lists unvalidated header fields, timing of permitted traffic and variation in response formatting as residual channels, and concludes that the passive tap must be replaced by an active one. ¹⁹

  Theoretical argument · Significant · Open. On the record

• Completeness rests on physical monitoring left out of scope in Network taps and certifiers

  The Secure Gateway Device paper assumes the facility is physically monitored, and states that the whole architecture depends on the device being the only communication channel. It names radio emanation, power-line signalling and thermal channels as covert channels beyond that scope. ¹⁵

  Open question · Significant · Open. On the record

  Related mechanism R1 Side-channel suppression for isolated facilities: Addresses the radio, power-line and thermal channels that network-level designs leave out. A pointer, not evidence that this flaw is mitigated. Add

• Prompt-controlled entropy inflation widens the covert channel in Bounding unexplained information in outputs

  Gumbel-based inference verification tolerates token choices that honest GPU nondeterminism could produce, and the size of that tolerated set grows with the model's output entropy. Kezins, an independent researcher, showed that an adversary who controls the prompt distribution can raise output entropy and roughly double the bits leaked per token. Across six models of 1 to 32 billion parameters, this cut the slowdown from 146–254 times under benign prompts to 60–118 times. Kezins argues that architectures built on the same unexplained-information bound inherit this attack surface, and recommends calibrating tolerances against local token entropy rather than benign traffic. ^{3 9}

  Demonstrated attack · Significant · Open. On the record

  Related mechanism R3 Deterministic and bit-exact inference: Bit-exact replay would remove the tolerance for numerical noise that sets the size of this channel. The record notes that it needs full hardware and software metadata. A pointer, not evidence that this flaw is mitigated. Add

• Information the declared computation explains is not bounded in Bounding unexplained information in outputs

  The bound limits unexplained bits only. Outputs that the declared computation fully explains can still carry valuable information: a compression study notes that an adversary with inference access can extract more proprietary information per bit than naive transmission allows. ^{24 25}

  Theoretical argument · Significant · Open. On the record

• Channels other than checked outputs are outside the bound in Bounding unexplained information in outputs

  The inference-verification scheme treats side channels as out of scope. A low-trust system design argues that suppressing physical covert bandwidth below kilobits per second is much more achievable than aiming for zero, and that a malicious device can leak one bit of information by deliberately outputting a wrong result. ^{3 17}

  Theoretical argument · Significant · Open. On the record

  Related mechanism R1 Side-channel suppression for isolated facilities: Physical side channels need separate suppression, which is this mechanism's purpose. A pointer, not evidence that this flaw is mitigated. Add

• The facility-level design is untested in Bounding unexplained information in outputs

  The compute-verification architecture is described with protocol details, potential attacks and prototyping plans, but no prototype results have been published. ²⁴

  Open question · Significant · Open. On the record

• Low-communication training reduces the bandwidth training needs in Bandwidth limits and compartmentalization

  DiLoCo matched fully synchronous training on 8 workers while communicating 500 times less. Rahman writes that this family of methods theoretically allows large-scale training with less than 100 Mbps. Lucid includes these methods in its bounds, but notes that extreme activation compression, architectures with unusually small inter-layer widths, or modular paradigms could erode the margin. ^{26 29 30}

  Theoretical argument · Significant · Open. On the record

• Operator control of pod routing collapses the bound in Bandwidth limits and compartmentalization

  Lucid's analysis finds that if the operator can freely assign pods to routers, it could dedicate a whole cell of 100 or more pods to one pipeline stage. The bound then falls to about 90–220x uncompressed and as low as about 25x with compression. The proposed mitigation, auditor-controlled random assignment that is periodically re-randomized, has not been implemented. ²⁶

  Theoretical argument · Significant · Open. On the record

• Undeclared local storage raises per-pod capacity in Bandwidth limits and compartmentalization

  More memory or storage per pod helps an adversary. Lucid requires per-pod storage to be declared, capped and physically inspected. ²⁶

  Theoretical argument · Significant · Open. On the record

• Training within one pod is not covered in Bandwidth limits and compartmentalization

  Lucid's bounds concern pre-training models larger than the pods are sized for. Training models that fit in one pod, fine-tuning and reinforcement-learning post-training within one pod are outside the modelled threat. ²⁶

  Open question · Significant · Open. On the record

• Parallel scale-up switches are hard enforcement points in Bandwidth limits and compartmentalization

  In GB200 topologies, GPUs reach GPUs in other nodes through NVSwitches without a NIC on the path. Amodo notes that limits are hard to enforce there because many switches work in parallel, so compromising one or two would bypass the limit. ²⁸

  Theoretical argument · Significant · Open. On the record

Not yet demonstrated

R1 Network taps and certifiers

Mechanisms on the map that are not in the proposal. Pointers, not recommendations. ?

• R1 Hardware-enabled guarantees (flexHEG) and guarantee processors Add
  • Bears on the open critical flaw “Underlying attestation can be forged or relayed” in Model identity attestation. A tamper-protected enclosure around the chip is the proposed answer when the party that holds the hardware may attack it physically.
• R3 Deterministic and bit-exact inference Add
  • Bears on the open significant flaw “Recomputation depends on trusted logging and randomness, and its tolerance leaves a covert channel” in Model identity attestation. Bit-exact inference would remove the numerical tolerance that leaves this channel.
  • Bears on the open significant flaw “Output nondeterminism leaves covert capacity” in Network taps and certifiers. Deterministic replay is one of the two remedies the flaw's source names.
  • Bears on the open significant flaw “Prompt-controlled entropy inflation widens the covert channel” in Bounding unexplained information in outputs. Bit-exact replay would remove the tolerance for numerical noise that sets the size of this channel. The record notes that it needs full hardware and software metadata.
  • Model identity attestation waits on it. Numerical nondeterminism limits how tightly recomputation can pin down the model and sampling.
  • Network taps and certifiers waits on it. Nondeterministic inference leaves covert capacity in outputs that hashing cannot remove.
  • Bounding unexplained information in outputs waits on it. Tolerance for numerical nondeterminism sets the size of the residual channel; bit-exact replay would remove it but needs full hardware and software metadata.
• R1 Side-channel suppression for isolated facilities Add
  • Bears on the open significant flaw “Completeness rests on physical monitoring left out of scope” in Network taps and certifiers. Addresses the radio, power-line and thermal channels that network-level designs leave out.
  • Bears on the open significant flaw “Channels other than checked outputs are outside the bound” in Bounding unexplained information in outputs. Physical side channels need separate suppression, which is this mechanism's purpose.
  • Network taps and certifiers waits on it. Radio, power-line and thermal channels are not addressed by network-level designs.
  • Bounding unexplained information in outputs waits on it. Physical side channels need separate suppression, and one design treats a low residual bandwidth, rather than zero, as the realistic target.
• R3 TEE remote attestation for AI workloads Add
  • Model identity attestation waits on it. Attestation that resists physical attackers, for the enclave variant.
• R2 Tamper evidence for verifier devices Add
  • Network taps and certifiers waits on it. Taps and gateway devices need tamper-evident housing and physical monitoring so that traffic cannot bypass them.
  • Bandwidth limits and compartmentalization waits on it. Shaping devices and routing assignments must be trusted by both parties; Amodo has not yet fully analysed resilience to a compromised DPU.
• R3 Sampled inference recomputation Add
  • Network taps and certifiers depends on it.
  • Bounding unexplained information in outputs depends on it.

Missing prerequisites

• R3 Sampled inference recomputation needed by Network taps and certifiers and Bounding unexplained information in outputs Add
• R3 Deterministic and bit-exact inference needed by Network taps and certifiers Add
• R2 Tamper evidence for verifier devices needed by Network taps and certifiers and Bandwidth limits and compartmentalization Add
• R1 Side-channel suppression for isolated facilities needed by Network taps and certifiers and Bounding unexplained information in outputs Add

Shared foundations

• Deterministic and bit-exact inference relied on by Model identity attestation, Network taps and certifiers and Bounding unexplained information in outputs
• Sampled inference recomputation relied on by Network taps and certifiers and Bounding unexplained information in outputs
• Tamper evidence for verifier devices relied on by Network taps and certifiers and Bandwidth limits and compartmentalization
• Side-channel suppression for isolated facilities relied on by Network taps and certifiers and Bounding unexplained information in outputs
• Network taps and certifiers relied on by Bounding unexplained information in outputs and Bandwidth limits and compartmentalization. In the proposal

Blockers

17 blockers recorded

• Model identity attestation
  • Attestation that resists physical attackers, for the enclave variant. Hardware trust. Waits on TEE remote attestation for AI workloads ⁸
  • Numerical nondeterminism limits how tightly recomputation can pin down the model and sampling. Protocol soundness. Waits on Deterministic and bit-exact inference ³
  • The recomputation variant needs the verifier to hold the declared weights. Access & governance ³
• Network taps and certifiers
  • No complete verification tap has been demonstrated at production frontend link rates, and on the tested CPU no hash algorithm reached line rate with minimum-size frames. Performance & compatibility ^{23 31}
  • Nondeterministic inference leaves covert capacity in outputs that hashing cannot remove. Evidence binding. Waits on Deterministic and bit-exact inference ¹⁵
  • Taps and gateway devices need tamper-evident housing and physical monitoring so that traffic cannot bypass them. Hardware trust. Waits on Tamper evidence for verifier devices ^{15 17}
  • Radio, power-line and thermal channels are not addressed by network-level designs. Coverage & hidden compute. Waits on Side-channel suppression for isolated facilities ¹⁵
  • Red-teaming by specialists is called for but has not been reported. Adversarial validation ¹⁵
• Bounding unexplained information in outputs
  • The prover's compute must be isolated so that all traffic passes through the verifier's interlock; any unmonitored path voids the bound. Coverage & hidden compute. Waits on Bandwidth limits and compartmentalization ²⁴
  • Physical side channels need separate suppression, and one design treats a low residual bandwidth, rather than zero, as the realistic target. Coverage & hidden compute. Waits on Side-channel suppression for isolated facilities ^{3 17}
  • Tolerance for numerical nondeterminism sets the size of the residual channel; bit-exact replay would remove it but needs full hardware and software metadata. Protocol soundness. Waits on Deterministic and bit-exact inference ^{9 17}
  • Recomputation over confidential weights and inputs needs a protected setting: prover recomputation in a verifier-controlled enclosure, verifier recomputation in a prover-controlled enclosure, or zero-knowledge proofs. Privacy & leakage ²⁴
  • No prototype of the facility-level architecture exists to red-team. Adversarial validation ²⁴
• Bandwidth limits and compartmentalization
  • No cap that a verifier can check has been implemented or red-teamed. Adversarial validation ²⁶
  • The verifier must know that all traffic leaving a pod crosses the capped, monitored links. Coverage & hidden compute. Waits on Network taps and certifiers ¹⁷
  • Shaping devices and routing assignments must be trusted by both parties; Amodo has not yet fully analysed resilience to a compromised DPU. Hardware trust. Waits on Tamper evidence for verifier devices ^{26 28}
  • Advances in low-communication training could shrink the margin that the cap enforces. Capacity bounds ^{26 29 30}

From the family or selected implementation's record. ?

Model weights

Depends on the design for Model identity attestation, Network taps and certifiers and Bounding unexplained information in outputs.

Not involved: Bandwidth limits and compartmentalization.

Inputs and outputs

Depends on the design for Model identity attestation, Network taps and certifiers and Bounding unexplained information in outputs.

Not involved: Bandwidth limits and compartmentalization.

Training data

Depends on the design for Network taps and certifiers.

Not involved: Model identity attestation, Bounding unexplained information in outputs and Bandwidth limits and compartmentalization.

Model identity attestation

• R2 Attestable Audits Research prototype, University of Cambridge
• R2 PAL*M Research prototype, University of Waterloo
• R3 Tinfoil model identity (Modelwrap) Product, Tinfoil

Network taps and certifiers

• R1 AI 2040 inference-only verification stack Proposed architecture, AI Futures Project
• R1 Low-trust AI compute verification system overview Proposed architecture, Machine Intelligence Research Institute
• R1 SASH confidential network logger Research prototype, Singapore AI Safety Hub (SASH)

Bounding unexplained information in outputs

None on the map

Bandwidth limits and compartmentalization

• R1 AI 2040 inference-only verification stack Proposed architecture, AI Futures Project
• R1 RAND secure inference data center (SIDC) design Proposed architecture, RAND

1. How Tinfoil Proves Exactly What Model Is Running, Tinfoil Team (2026). Original
2. PAL*M: Property Attestation for Large Generative Models, P. Chantasantitam et al. (2026). Original
3. Verifying LLM Inference to Detect Model Weight Exfiltration, R. Rinberg et al. (2025). Original
4. A primer on secure enclaves, Tinfoil (2026). Original
5. Backend infrastructure, Tinfoil (2026). Original
6. How verification works in Tinfoil, Tinfoil (2026). Original
7. modelwrap: Reproducible dm-verity read-only image of Huggingface models, Tinfoil (2026). Original
8. TEE.fail: Breaking Trusted Execution Environments via DDR5 Memory Bus Interposition, J. Chuang et al. (2026). Original
9. Adversarial Entropy Inflation Against Gumbel-Based Inference Verification, N. Kezins (2026). Original
10. Battering RAM: Low-Cost Interposer Attacks on Confidential Computing via Dynamic Memory Aliasing, J. De Meulemeester et al. (2026). Original
11. RMPocalypse: How a Catch-22 Breaks AMD SEV-SNP, B. Schlüter & S. Shinde (2025). Original
12. SEV-SNP RMP Initialization Vulnerability (AMD-SB-3020), AMD (2025). Original
13. On TEEs for Privacy-Preserving Monitoring in AI Governance, Gloria Z (2026). Original
14. Attestable Audits: Verifiable AI Safety Benchmarks Using Trusted Execution Environments, C. Schnabl et al. (2025). Original
15. Fingerprinting All AI Cluster I/O Without Mutually Trusted Processors, N. Cankaya et al. (2026). Original
16. The Fundamentals and Feasibility of Secure Network Taps for Verifying AI Datacenter Use, N. Cankaya (2026). Original
17. A System Overview for Near-Term, Low-Trust AI Compute Verification, N. Cankaya (2026). Original
18. Verification Plan, R. Dean (2026). Original
19. Fitting a Network TAP to our Inference Verification Prototype, Amodo Design (2026). Original
20. Amodo-Design/Inference-Recomputation-Prototype (GitHub repository), Amodo Design (2026). Original
21. inference-verification: Inference Verification Prototype, Singapore AI Safety Hub (SASH) (2026). Original
22. Internationalising AI Verification, Singapore AI Safety Hub (SASH) (2026). Original
23. Network Tapping for AI Verification: A Technical Assessment, Amodo Design (2026). Original
24. Verifying AI Compute by Bounding Unexplained Information Exfiltration, J. Petrie & Y. Mühlhäuser (2026). Original
25. Haiku to Opus in Just 10 bits: LLMs Unlock Large Compression Gains, R. Rinberg et al. (2026). Original
26. Traffic Shaping for Workload Classification, Lucid Computing (2026). Original
27. De-risking Interconnect Limits for AI Verification, A. Scher et al. (2026). Original
28. The Tray as a Bandwidth Boundary, Amodo Design (2026). Original
29. DiLoCo: Distributed Low-Communication Training of Language Models, A. Douillard et al. (2024). Original
30. Does Distributed Training Undermine Compute Governance?, R. Rahman (2026). Original
31. Network Traffic Hashing, Amodo Design (2026). Original