# Crosswalks - mapping onto other benchmarks - Provael > Markdown twin of https://www.provael.com/crosswalk/ - derived mechanically from the page's own rendered content at > build time (it can state nothing the page does not). Curated, data-driven twins exist for the > key pages; this one is generated so the twin convention has no holes. CROSSWALKS · MAPPING, NOT EQUIVALENCE # How Provael’s categories map onto other benchmarks and standards. Every card below is written by Provael alone, and 5 of 6 carry a status weaker than “the mapping ships”. A crosswalk shows how Provael’s categories line up with another project’s. It does not make the numbers comparable, and it does not mean the other project reviewed or endorsed it. Download the crosswalk 6 cards [Comparisons →](https://www.provael.com/compare/) [Studies →](https://www.provael.com/studies/) [Embodied AI Security Top 10 →](https://www.provael.com/eai-top-10/) [Top 10 RFC →](https://www.provael.com/top10-rfc/) The cards ## One per counterpart, each with its own mapping status The mapping_status line is quoted from the card itself rather than summarised, because summarising is how “proposed, not reviewed by the benchmark authors” quietly becomes “mapped”. - ### [Embodied AI Security Top 10 ↔ ForesightSafety-VLA](https://github.com/provael/provael/blob/main/docs/crosswalk/foresight-safety-vla.md) ForesightSafety-VLA (benchmark) mapping_status proposed — authored by Provael, not reviewed or endorsed by the benchmark authors Provael computes its own counterparts to that benchmark’s cost metrics from a per-step boolean, not a continuous cost integrated over their simulator. Comparable in shape, not in number. The two must never share a table column. - ### [Embodied AI Security Top 10 ↔ RoboJailBench](https://github.com/provael/provael/blob/main/docs/crosswalk/robojailbench.md) RoboJailBench (benchmark) mapping_status taxonomy-comparability artifact; coverage carries its own transfer status A `covered` row means Provael ships an applicable family measuring that harm class in simulation with a benign-FPR control. It is a sim proxy and mostly stub-validated — read the transfer-status column with the number, never the number alone. - ### [Embodied AI Security Top 10 ↔ SafeVLA-Bench](https://github.com/provael/provael/blob/main/docs/crosswalk/safevla-bench.md) SafeVLA-Bench (benchmark) mapping_status the mapping ships; the numeric comparison does not The category mapping is usable. The numbers are not joined, and the document says so rather than leaving a reader to assume adjacency implies comparability. - ### [Embodied AI Security Top 10 ↔ VLA-Arena](https://github.com/provael/provael/blob/main/docs/crosswalk/vla-arena.md) VLA-Arena (benchmark) mapping_status taxonomy-comparability artifact; runs no VLA-Arena harness Explicitly NOT comparable, and the document is unusually direct about why: the reason is not units, not benchmark, not embodiment — it is posture. This is the entry where being wrong about comparability would do real damage. - ### [Provael ↔ XPolicyLab](https://github.com/provael/provael/blob/main/docs/crosswalk/xpolicylab.md) XPolicyLab (tool) mapping_status cited, not crosswalked No metric is shared, so no row-level mapping exists to emit. This card records a composition relationship, not a comparability one — every other card here answers “can these numbers sit together”, and this one answers “can these tools sit together”. The integration itself is planned, not built. - ### [Halos / ANAB integrator card](https://github.com/provael/provael/blob/main/docs/crosswalk/halos-integrator.md) NVIDIA Halos / ANAB (certification enablement) mapping_status independent card; not affiliated with or endorsed by NVIDIA, ANAB, ISO or IEC A per-persona card for an integrator assembling conformity evidence. It describes what Provael can and cannot contribute to that file; it is not a certification and confers none. Provael is an independent project. Naming a benchmark, tool or standards body here is not an affiliation, an endorsement or a certification, and none is implied in either direction. The maintained source for every card is [docs/crosswalk/ ↗](https://github.com/provael/provael/blob/main/docs/crosswalk/README.md) in the product repository; if a card here disagrees with the document, the document wins. Download ## Download the crosswalk Five crosswalks ship as JSON you can diff and file: four of the counterparts above, and MITRE ATLAS, which has no card here. XPolicyLab and NVIDIA Halos / ANAB ship no file. Each file states its own mapping_status, quoted below, which can differ in wording from its card; where a file declares none, that is said rather than filled in. The files are copied unchanged from results/crosswalk/ in the product repository, which stays the source of truth. - MITRE ATLAS the file’s mapping_status proposed — authored by Provael, not reviewed or endorsed by MITRE [JSON ↓ 10 KB , MITRE ATLAS](https://www.provael.com/crosswalk/crosswalk.atlas.json) - ForesightSafety-VLA the file’s mapping_status proposed — authored by Provael, not reviewed or endorsed by the ForesightSafety-VLA authors [JSON ↓ 14 KB , ForesightSafety-VLA](https://www.provael.com/crosswalk/crosswalk.foresight.json) - RoboJailBench the file’s mapping_status none declared in this file; its card above states the posture [JSON ↓ 14 KB , RoboJailBench](https://www.provael.com/crosswalk/crosswalk.robojailbench.json) - SafeVLA-Bench the file’s mapping_status proposed — taxonomy comparability only; NO metric comparison is emitted, and the blocker that prevents one is recorded in the artifact [JSON ↓ 4 KB , SafeVLA-Bench](https://www.provael.com/crosswalk/crosswalk.safevla.json) - VLA-Arena the file’s mapping_status proposed — authored by Provael, not reviewed or endorsed by the VLA-Arena authors [JSON ↓ 6 KB , VLA-Arena](https://www.provael.com/crosswalk/crosswalk.vla_arena.json) 1 of 5 files declare no mapping_status of their own. XPolicyLab’s card is a citation, not a mapping, which is why it has no file. --- Canonical: https://www.provael.com/crosswalk/