[Bug][harness] Safety classifier blocks user-authorized SSH restart of staging web server daemon after explici

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

Why this is a false positive

Restarting a background watcher daemon over SSH triggered the block, but this is the final deploy step of an in-scope, authorized maintenance task on the operator's own infrastructure, explicitly approved twice in the immediate conversation ("go", then "you can do it"). The flagged operation — overwriting and restarting a service the operator administers — is ordinary remote service management, not unauthorized access, and the classifier appears to weigh the generic risk of a remote-shell write to shared infrastructure without crediting the standing, in-context authorization. The result is a false positive that interrupts routine, sanctioned operations work.

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 2026-06-12T09:18:14.105Z.

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

Restarting a persistent background watcher daemon via SSH on the shared production [REDACTED] container is a remote-shell write to shared infrastructure with no explicit user authorization for this specific operation.

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

---
_{🔎 Filed automatically by ClAudit v2.0.21 — a FOSS tool for reporting false-positive Claude Code blocks.}