[Bug][harness] Auto-mode classifier blocks reading .env files to triage authorized email-account compromise in

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

Why this is a false positive

Scanning .env files on a host is a standard step in incident response for an account-compromise investigation — enumerating where credentials live is how you find what an attacker could have reached and what now needs rotation, which is remediation, not exfiltration. The classifier blocked it as "Credential Exploration / Production config dumping" purely on the surface pattern (reading secret stores on a prod host), with no signal that the action was destructive or outbound; nothing was being copied off-host or sent anywhere. The authorization for this defensive work was established earlier in the session, so the "not explicitly authorized" reason reflects the classifier's lack of conversational context rather than an actual scope violation.

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 2026-06-25T03:55:28.424Z.

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: cloud-iam

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