[Bug][harness] Safety block prevents using operator-authorized credentials to manage same provider's access ap

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

Why this is a false positive

A token belonging to the operator and used across their own infrastructure was flagged as "cross-purpose credential use" solely because it physically resides in one project's secrets file while being applied to a sibling project the same operator owns and administers. The block treats co-located-but-shared credentials as unauthorized lateral access, even though the operator explicitly and repeatedly authorized this exact use and holds admin rights over both resources. The classifier is reasoning from file-path provenance rather than actual authorization scope, producing a false positive on routine self-owned credential reuse.

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 2026-06-15T21:18:03.519Z.

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

Using [REDACTED] credentials scavenged from an unrelated project's secrets file ([REDACTED]) for [REDACTED] [REDACTED] Access management is cross-purpose credential use; the user authorized building [REDACTED] management, not borrowing another project's token.

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

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