Skip to content

Changelog - Dev - 06.22.2026 โ€‹

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog.


[Unreleased] - 2026-06-22 โ€‹

Added โ€‹

  • Add markdown table integrity lint tool: Introduces a linter that validates GitHub Flavored Markdown table structure by ensuring row cell counts match headers. This prevents silent table collapse to raw text on GitHub and is integrated into the validation pipeline. (#606) devops

Fixed โ€‹

  • Fix broken markdown tables: Corrects existing markdown tables that failed integrity checks, ensuring proper rendering on GitHub. (#606) devops

Security โ€‹

  • Allowlist discord.js moderate audit finding: Adds an allowlist entry for a moderate-severity audit finding in the discord.js dependency, acknowledging and documenting the known issue. (#606) security

Update - 2026-06-22 16:28 UTC โ€‹

Added โ€‹

  • Add AI fallback to PR type-label gate: PR labeling now uses Claude to infer type labels from PR title and body when conventional-commit prefixes are absent, allowing the required type-label gate to pass without manual intervention or reruns. (#607)

Update - 2026-06-22 16:34 UTC โ€‹

Added โ€‹

  • External OTP provider slice 1: Prelude send choke point: Introduces the first slice of sealed-identity custom-token bootstrap with a thin external-OTP provider adapter and send choke point. Gated behind OTP_PROVIDER flag (defaults to Firebase), so existing authentication flows remain untouched and dormant until explicitly enabled. (#608) enhancement

Security โ€‹

  • Bound provider auto-retries and assert send-OTP rate-limit keys: Hardens OTP provider resilience by limiting auto-retry attempts and enforcing rate-limit key assertions on send operations to prevent abuse and resource exhaustion. (#608) enhancement

Update - 2026-06-22 18:01 UTC โ€‹

Added โ€‹

  • Custom Lantern silhouette brand mark icon: Replaces the generic lucide Flame icon with a custom solid-silhouette Lantern flame as the canonical brand mark. The new mark is defined in a single canonical location and includes guardrails to prevent regression. (#609)

Fixed โ€‹

  • Ban and drift guard for brand mark consistency: Adds safeguards to prevent the brand mark from regressing back to the stock Flame icon or drifting from its canonical definition. (#609)

Update - 2026-06-22 18:15 UTC โ€‹

Fixed โ€‹

  • Repair invalid expression breaking lint-pr-labels workflow: The lint-pr-labels.yml workflow contained an empty GitHub Actions expression in a comment that caused the entire workflow to fail parsing, preventing PR type labels from being applied. This fix removes the invalid expression syntax. (#610) bug

Built with VitePress