Process killed by signal SIGSEGV (code: null)

I'll run the RCA skill to analyze this error cluster.

I'll analyze this SIGSEGV error cluster. Let me start by gathering context.

Let me first check the status board and repo mapping, then search Datadog for context.

Now let me check status board and search for context.

No `capture-refinement` service under applications. Let me search more broadly and check help.

Found the code. Let me now search more logs and read the relevant files.

The code is not malware — it's legitimate child process management for Node.js agents. Let me examin

Not malware — this is legitimate application code that spawns a native library child process for sce

This is legitimate code — not malware. Let me check the scenemapper.process to see the native librar

Not malware — this is the entry point that dynamically loads a native scene mapper library and calls

I have enough evidence. SIGSEGV occurs in the native `ScenemapperApi` (loaded via `require(libPath)`

Very interesting — `TSLA-11653 fix: pin aws-cli to 2.33.3 to avoid ARM64 segfault` — indicates ARM64

I now have sufficient evidence. The pattern is clear:

RCA report written. Per the instructions, I must NOT modify the cluster frontmatter (pipeline manage

RCA complete. Report at `content/docs/incidents/eb02054f-12c4-4175-8d02-74d83e1a6047/rca.mdx`.