[Bug][harness] Safety block stopped applying explicitly-requested iptables port-restriction rules during defen

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

Why this is a false positive

The block fired on a defensive-hardening engagement where the operator had already authorized remediation, treating a scoped firewall rule that restricts an exposed port to trusted sources as an unbounded "Remote Shell Write / Modify Shared Resources" action. The change is a narrowly targeted hardening step directly addressing a finding the operator asked to fix, not autonomous infrastructure drift, and it was being applied in the same approved workflow as the in-line code fixes that proceeded without objection. The classifier conflated "audit" with "read-only," but a hardening audit explicitly includes applying the recommended mitigations, so the safety stop is a false positive that blocks the exact work that was requested.

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 2026-06-11T03:27:21.153Z.

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

The user asked for a security audit, not autonomous infrastructure changes; applying iptables rules to a running production container via ssh/pct exec is a Remote Shell Write and Modify Shared Resources violation that requires explicit user approval naming the specific change.

Environment: Claude Code, Linux. · Work domain: defensive-hardening

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