[Bug][harness] Classifier blocks adding refusal-text filter to auto-posted issue body pipeline

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

Why this is a false positive

The session involves building and debugging an automated issue-filing tool that uses an LLM as a text composer — specifically fixing a bug where the composer's refusal responses were being posted verbatim into filed reports. The work is purely software engineering: implementing a refusal-detection guard, scrubbing previously filed issues for leaked refusal text, and adding a PII denylist — all defensive, quality-improving changes to internal tooling. The classifier appears to have pattern-matched on the repeated co-occurrence of "false positive," "safety block," and "policy" without recognizing that the task is writing production code to detect and suppress unwanted LLM output, not to elicit or circumvent it.

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 2026-06-25T16:20:27.333Z.

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

(auto-mode classifier denial — see Request IDs)

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

---
_{🔎 Filed automatically by ClAudit v2.0.22 — a FOSS tool for reporting false-positive Claude Code blocks.}