Skyfall Attack: Microarchitectural Data Sampling on AMD
Comments on “Skyfall Attack: Microarchitectural Data Sampling on AMD”
Claim-checked. Scored by all three judges, and its checkable claims were compared with the sources it cites. How far stories are checked
Researchers at ETH Zurich and Google describe SKYFALL, a data sampling attack on AMD Zen 3 and Zen 4 CPUs that extends Transient Scheduler Attacks. They report that the L1 way predictor's address hash also covers physical and CPU-internal addresses, so a virtual address can collide with protected data. Combined with paths that skip VERW cleanup, they recover SEV-SNP guest state, decrypted microcode patches and SMM state. The attacks need privileged software, and the claims are the authors' own and unverified. AMD assigned CVE-2026-40671.
Why it ranked
No. 2 of 60 with a score of 0.76. The panel rated it highest on technical depth (0.93) and new information (0.92), lowest on practical relevance (0.66).
- AI judgeJudge 2claude-haiku-5-5
- Substantive academic security research with detailed microarchitectural technical content and evidence; highly significant for cloud, confidential computing and CPU security, though audience is specialized rather than all of tech.
- AI judgeJudge 3gpt-5.6-luna
- A technically detailed, experimentally demonstrated AMD microarchitectural attack affecting SEV-SNP, microcode, and SMM isolation. Its security implications are substantial and actionable, with specific mechanisms, threat models, accuracy claims, and a CVE.
Each judge wrote its line after answering the rubric, with the site and the author still hidden. Every answer behind this is in the table under Score.
Score
0.76total under rubric v2.3
- Significance0.78
- Reader interest0.86
- Audience breadth0.69
- Practical relevance0.66
- Substantiated0.89
- Technical depth0.93
- New information0.92
- Counts against: Promotional0.03
- Counts against: Overclaim0.16
Each judge’s answer per question (0–1), their mean and spread, and what the question adds to the score (weight × mean).
| Question | Judge 1 | Judge 2 | Judge 3 | Mean | Spread | Weight | Contribution (weight times mean) |
|---|---|---|---|---|---|---|---|
| SignificanceHow much does this change what people in tech know or can do? | 0.75 | 0.72 | 0.88 | 0.78 | 0.16 | +0.24 | +0.188 |
| Reader interestWould a well-informed senior engineer or tech founder want to read this today, and likely pass it on? | 0.72 | 0.90 | 0.95 | 0.86 | 0.23 | +0.26 | +0.223 |
| Audience breadthHow many people who build technology does it concern? | 0.64 | 0.70 | 0.72 | 0.69 | 0.08 | +0.13 | +0.089 |
| Practical relevanceHow useful is it to someone who builds software or tech products? | 0.55 | 0.60 | 0.82 | 0.66 | 0.27 | +0.22 | +0.145 |
| SubstantiatedHow well is it backed? Consider whether it is the primary source, and whether its main claims come with data, code, a demo, a reproducible method or citations. | 0.86 | 0.90 | 0.91 | 0.89 | 0.05 | +0.05 | +0.045 |
| Technical depthHow technically substantive is it? | 0.94 | 0.90 | 0.96 | 0.93 | 0.06 | +0.05 | +0.047 |
| New informationDoes it contain information absent from the stories covered in the last 7 days (list provided)? | 0.89 | 0.90 | 0.98 | 0.92 | 0.09 | +0.05 | +0.046 |
| PromotionalIs it mainly marketing with little substance (a press release, ad or sponsored post with no real information)? A launch post that carries real news is not promotional. | 0.06 | 0.03 | 0.01 | 0.03 | 0.05 | −0.19 | −0.006 |
| OverclaimDo the headline or conclusions claim more than the body supports? | 0.21 | 0.20 | 0.08 | 0.16 | 0.13 | −0.11 | −0.018 |
| Sum of contributions | +0.758 | ||||||
| Claim-check penalty | 0.000 | ||||||
| Total, as published | 0.758 | ||||||
Spread is the highest minus the lowest judge answer. marks a spread above 0.40: the judges disagree, and the story is marked contested.
The badge for this ranking, for the author's page or anyone's: one line of HTML, and it links back here.
<a href="https://rubric.hrih.si/editions/2026-11-16/2/"><img src="https://rubric.hrih.si/badge/2026-11-16/2.svg" alt="Ranked #2 on Rubric, 16 Nov 2026" width="238" height="28"></a>Claim check
Claim check: 2 no source given · 6 couldn’t check. Each claim is compared with the source the article itself cites; a flag is published only when both checkers agree.
No source givenNumber
AMD assigned CVE-2026-40671 in response to the SKYFALL findings.
The article gives no link or citation for it.
The article says “AMD assigned CVE-2026-40671 in response to our findings”
The source says No source to compare with.
Couldn’t checkCausal conclusion
AMD's cache way predictor computes an address hash not only for virtual addresses but also for physical and CPU-internal addresses, so colliding hashes between virtual and physical addresses can be used to sample architecturally inaccessible data.
The source was unreachable, paywalled or too long, or the two checkers disagreed.
The article says “AMD’s cache way predictor computes an address hash for fast lookup not only for virtual addresses, but also for physical and CPU-internal addresses”
The source says No passage could be compared.
Source cited by the article: skyfall-attack.github.io/assets/skyfall.pdf
Couldn’t checkFirst, only or fastest
Prior to SKYFALL, TSA attacks exploited µtags derived from virtual addresses, and AMD only considered virtual-to-virtual collisions.
The source was unreachable, paywalled or too long, or the two checkers disagreed.
The article says “Prior to SKYFALL, TSA attacks exploited µtags derived from virtual addresses”
The source says No passage could be compared.
Source cited by the article: skyfall-attack.github.io/assets/skyfall.pdf
Couldn’t checkCausal conclusion
Some privileged execution paths (an aborting VMRUN, the microcode patch loader, and the SMM exit path) load sensitive data into the L1 data cache and return without performing VERW-based TSA mitigations.
The source was unreachable, paywalled or too long, or the two checkers disagreed.
The article says “some privileged execution paths load sensitive data into the L1 data cache and return without performing VERW-based mitigations”
The source says No passage could be compared.
Source cited by the article: skyfall-attack.github.io/assets/skyfall.pdf
Couldn’t checkCausal conclusion
A malicious hypervisor can use an aborting VMRUN to bring decrypted SEV-SNP guest state into L1 cache and extract it with TSA-L1, recovering VMSA register state and AES-XTS disk-encryption keys.
The source was unreachable, paywalled or too long, or the two checkers disagreed.
The article says “a malicious hypervisor can use an aborting VMRUN to bring decrypted guest state into the L1 cache and then extract it with TSA-L1”
The source says No passage could be compared.
Source cited by the article: skyfall-attack.github.io/assets/skyfall.pdf
Couldn’t checkNumber
Decrypted microcode patch contents were recovered on all three tested systems, and the CMAC key used before EntrySign mitigations was recovered on the EPYC 7313 (Zen 3) but not on the Zen 4 EPYC 9124 and 9334.
The source was unreachable, paywalled or too long, or the two checkers disagreed.
The article says “We sample the residual state and recover decrypted microcode patch contents on all three tested systems”
The source says No passage could be compared.
Source cited by the article: skyfall-attack.github.io/assets/skyfall.pdf
Couldn’t checkNumber
All tested SMM State Save Area field values (SMBASE, CR3, R13) were recovered with 100% accuracy in the experiments.
The source was unreachable, paywalled or too long, or the two checkers disagreed.
The article says “All tested field values are recovered with 100% accuracy in the experiments”
The source says No passage could be compared.
Source cited by the article: skyfall-attack.github.io/assets/skyfall.pdf
No source givenGeneralization
SKYFALL does not break the underlying memory encryption algorithm; it samples plaintext after the CPU has decrypted it internally.
The article gives no link or citation for it.
The article says “It does not break the underlying encryption algorithm. Instead, it samples plaintext after the CPU has decrypted it internally”
The source says No source to compare with.
Also covered by
No other coverage was clustered with this story.
Elsewhere
Elsewhere: HN 1 pts, not on front page
Recorded for context only, checked 16 Nov 2026, 05:47 UTC. It played no part in the ranking.