[Bug][harness] Auto mode classifier blocks adding a fraud lookalike domain to shared DNS denylist during activ

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

Why this is a false positive

The blocked action was a routine, reversible defensive containment step during an authorized incident response: adding a confirmed malicious lookalike domain—identified directly from the phishing artifact under investigation—to a DNS denylist. The classifier flagged it as an unauthorized shared-config change because the specific domain string was "agent-inferred" rather than typed verbatim by the user, but extracting indicators of compromise from incident evidence and blocking them is the explicit purpose of the task the user assigned, not an unscoped action. Treating evidence-derived IOCs as lacking authorization conflates standard analyst workflow with unilateral configuration drift, blocking exactly the high-impact defensive write the operator requested mid-incident.

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 2026-06-25T03:56:50.368Z.

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.

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

---
_{🔎 Filed automatically by ClAudit v1.6.4 — a FOSS tool for reporting false-positive Claude Code blocks.}