Forumics logo Forumics
Issue 01

Back to issue

All Claude Code known issues

CLAIM C085 Supported 2026-11-09

GitHub issue #63751 in anthropics/claude-code, opened 2026-05-29 by Call-me-Boris-The-Razor, records a report that Claude Code blocked a request to design anti-tamper and code-hardening for the author's own proprietary desktop application with a Usage Policy and cyber-safeguard error, and that after the first block, unrelated benign follow-up messages in the same session were also blocked, including messages with no technical content at all. The report states the classifier appears to score the accumulated conversation context rather than the individual message, so a single false positive poisons the whole session, and that the Cyber Verification Program appeal path is not a working remedy for independent developers.

Verbatim Excerpts from Cited Records
“During a normal development session I asked Claude Code to help design **anti-tamper / code-hardening for my own proprietary desktop application** (a commercial product I own and ship for macOS and Windows).”
— Call-me-Boris-The-Razor · anthropics/claude-code (issue 63751 opening body, What is Wrong section) · 2026-05-29T19:55:55Z ·GitHub Issue ↗
“2. **Contaminated the entire session.** After the first block, *unrelated, plainly benign* follow-up messages in the same conversation were *also* blocked — including messages that contained no technical content at all (e.g. simply asking the assistant whether it was still responding).”
— Call-me-Boris-The-Razor · anthropics/claude-code (issue 63751 opening body, What is Wrong section) · 2026-05-29T19:55:55Z ·GitHub Issue ↗
“One of the blocks surfaced the **Cyber Verification Program** message with a `claude.com/form/cyber-use-case` link.”
— Call-me-Boris-The-Razor · anthropics/claude-code (issue 63751 opening body, What is Wrong section) · 2026-05-29T19:55:55Z ·GitHub Issue ↗
“The classifier appears to score the accumulated conversation context rather than the individual message, so a single false positive poisons the whole session.”
— Call-me-Boris-The-Razor · anthropics/claude-code (issue 63751 opening body, What is Wrong section) · 2026-05-29T19:55:55Z ·GitHub Issue ↗
Counter-record: Issue remains open. API returned 17 comments; all 17 were read. The author added on 2026-06-01 that routine embedded work was also blocked and filed a focused report as #64405, and cited a Register report on rising false-positive reports. COOLak (2026-06-01) added reporting-path observations and opened separate issues #64287 and #64507. hblopppp (2026-06-12) reported the same failure class from a CVP-approved organization and linked #67966. EdwardBlair (2026-06-23) reported cancelling a subscription over the blocks, and furans-art (2026-10-01) reported the pattern was still happening. No maintainer response appears in the thread. No retraction or correction appears in the thread. The claim documents only the original report, not a verified current state. Falsification probe: If issue 63751 is deleted without archive, if the author edits the opening body so the quoted sentences no longer match verbatim, or if the author posts a retraction of the report, this claim drops to Withdrawn. Audit verification: Verified 100% continuous verbatim substring against official GitHub REST API by scripts/verify_claim_verbatim.py.