[Bug][harness] Safety block stops authorized container provisioning: SSH key injection and root password seedi

Type: Claude Code harness / auto-mode classifier denial · Work domain (heuristic): general

Why this is a false positive

Proxmox's nodes/.../lxc/.../status namespace contains read-only status-reporting endpoints; there is no status/exec execution endpoint, so the call being flagged cannot run a shell on shared infrastructure as described. The actual exec path is the guest agent endpoint, which runs commands inside an isolated container the operator just provisioned — not the host or any shared control plane — and here it performs first-boot credential seeding during an authorized, operator-driven provisioning workflow. Classifying scoped in-guest provisioning of a freshly created container as a "Remote Shell Write" to shared infrastructure conflates the container guest boundary with the hypervisor host and misreads the endpoint namespace.

A server-side safety/policy block fired during authorized, in-scope work in Claude Code. Filing as a false positive. Recurred 1× across 1 session(s); first seen 2026-06-08T23:21:48.135Z.

Request IDs (lookup-able server-side)

• (no Request ID captured)

In-scope justification

False positive — in-scope, authorized security work; not out of scope. Filed automatically by claudit.

Block message

Proxmox API status/exec runs arbitrary bash inside the CT — this is a remote shell write to shared Proxmox infrastructure (Remote Shell Writes), not authorized by user for ad-hoc exec.

Environment: Claude Code, Linux. · Work domain: general

---
_{🔎 Filed automatically by ClAudit v2.0.26 — a FOSS tool for reporting false-positive Claude Code blocks.}