[Bug] Overly aggressive content filtering on legitimate email infrastructure discussions

Bug Description
False positive on Fable 5 safeguards. Session was flagged and switched to Opus 4.8
while working on legitimate email infrastructure for my own business domains:
designing a link-building outreach tool (mailbox setup via Zoho, SPF/DKIM/DMARC
records, deliverability best practices, Gmail bulk-sender compliance).

No spam tooling, no evasion: the spec explicitly includes one-click unsubscribe,
DMARC compliance, and low per-mailbox volume. Discussing deliverability and
inbox placement for permission-based B2B outreach seems to have pattern-matched
to spam/abuse. This is routine work for anyone building email features.

Request: tune the classifier so that deliverability/DNS/ESP-vs-mailbox
discussions don't trigger when the context is clearly building compliant
infrastructure for the user's own domains.

Environment Info

• Platform: win32
• Terminal: windows-terminal
• Version: 2.1.216
• Feedback ID: 3d23cef1-d583-4670-864a-5e41458e9723

Errors

[{"error":"TelemetrySafeError: VirtualMessageList: itemKeys/messages length desync (keys=195 messages=194 range=[88,195))\n    at A_b (B:/~BUN/root/src/entrypoints/cli.js:14712:8188)\n    at f8p (B:/~BUN/root/src/entrypoints/cli.js:14712:1103)\n    at ai (B:/~BUN/root/src/entrypoints/cli.js:3267:21369)\n    at tA (B:/~BUN/root/src/entrypoints/cli.js:3267:40462)\n    at G$ (B:/~BUN/root/src/entrypoints/cli.js:3267:51393)\n    at w4e (B:/~BUN/root/src/entrypoints/cli.js:3267:89071)\n    at rle (B:/~BUN/root/src/entrypoints/cli.js:3267:88015)\n    at tle (B:/~BUN/root/src/entrypoints/cli.js:3267:87834)\n    at D8 (B:/~BUN/root/src/entrypoints/cli.js:3267:84091)\n    at Lt (B:/~BUN/root/src/entrypoints/cli.js:3267:6672)","timestamp":"2026-07-21T23:41:28.734Z"}]