Release archive
3.2.158

CRYPTO-PDF-AI-PROMPT-SCORE-PERCENT-GUARD (Narrative Safety)

Release v3.2.158 focuses on CRYPTO-PDF-AI-PROMPT-SCORE-PERCENT-GUARD (Narrative Safety).

FixedInternal
CRYPTO-PDF-AI-PROMPT-SCORE-PERCENT-GUARD (Narrative Safety)

Release digest

Fixed

  • CRYPTO-PDF-AI-PROMPT-SCORE-PERCENT-GUARD (Narrative Safety) — Crypto PDF AI prompt inputs now normalize score and percent-style fields through finite numeric formatters so Trust Score, team score, security score, history trend, and unlock progress do not render as nan/100, inf/100, not scored/100, or inf% unlocked.

Internal

  • Added prompt-contract coverage proving non-finite score and unlock-progress inputs render as missing evidence copy instead of factual score or percent evidence.
  • Updated docs/architecture/CRYPTO_REPORTS.md because this tightens the internal AI prompt/report contract. Public-docs decision: no user-facing docs change required because downloadable PDF sections and dashboard behavior are unchanged.

Technical notes