Skip to content

[go-LIN-N2-01] HARA ASIL ratings appear one severity band too high for every hazard #62

Description

@SoundMatt

Location

.fusa-hara.json hazard entries H-01 through H-06 (around lines 27, 34, 41, 48, 55, 62).

Finding

Checking each hazard's declared asil against the standard ISO 26262 Severity/Exposure/Controllability determination table, the recorded ASIL looks one band too conservative-high across the board:

  • H-01 through H-04 all carry S2/E3/C2, which the standard table resolves to ASIL-A, but the JSON declares ASIL-B for each.
  • H-05 carries S1/E3/C2, which resolves to QM, but the JSON declares ASIL-A.
  • H-06 carries S2/E2/C2, which also resolves to QM, but the JSON declares ASIL-A.

Since this offset is uniform across every hazard, it looks like a systematic banding error in whatever generated the HARA rather than six independent judgment calls. Safety goals SG-01, SG-02, and SG-05 inherit the inflated ASIL-B from H-01/H-03 (and H-02/H-04).

I independently cross-checked the S/E/C combinations against the published ISO 26262-3 ASIL determination table and the discrepancy holds for all six hazards, so this is not just an artifact of a naive summation heuristic — it also disagrees with the real standard table.

Recommendation

Recompute each hazard's asil field directly from its S/E/C combination against the ISO 26262 determination table, or, if the higher ratings are an intentional conservative up-rating, state that rationale explicitly in the HARA rather than presenting it as the raw computed determination.

Filed from the 2026-07-29 ecosystem audit register; independently re-verified against current HEAD before filing.

Metadata

Metadata

Assignees

No one assigned

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions