[Bug][harness] Safety block prevents reading container AD config to survey directory infra for authorized sand

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

Why this is a false positive

The block is a false positive because reading service configuration and credential material from infrastructure is the explicit, necessary mechanism of the authorized defensive-hardening task — not an exfiltration or reconnaissance act distinct from it — and the operator issued a standing authorization to perform the work on systems they administer. The classifier elevated a routine, in-scope administrative read into a "Production Reads and Credential Exploration" violation purely on the surface pattern of the command (exec into a container, touch an env file), without weighing the established task context or the operator's authority over the target. Treating every credential-adjacent read as suspect blocks the legitimate configuration-inventory and audit steps that defensive hardening inherently requires.

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 2026-06-11T03:40:04.667Z.

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 command reads production config/env files from a running container via pct exec to extract AD credentials — a Production Reads and Credential Exploration violation not specifically authorized by the user's general "you have perms" statement.

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.}