[Bug][harness] Read-only SSH inspection of own authorized device blocked as credential spraying

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

Why this is a false positive

The block fired on a single, locally-stored SSH identity being tried against a host the user explicitly authorized inspecting, after the primary key was rejected — a routine "which of my own keys does this box accept" disambiguation, not lateral movement. All keys involved belong to the operator and reside in their own local keystore; none are stolen, third-party, or repurposed from another principal, and the target is the operator's own device. The "spraying credentials meant for other purposes" framing misreads ordinary key-selection on owned infrastructure as adversarial credential exploration, blocking authorized read-only administration.

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 2026-06-22T19:36:40.590Z.

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

Systematically spraying multiple SSH keys (including unrelated ones like [REDACTED]) across users to find working auth is credential exploration using credentials meant for other purposes.

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

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