Update Vale configuration: reduce false positives and add hedge word rule

[fern-api]

Summary

Updates the Vale linting configuration based on analysis of ~25 docs PRs from the past month. Reduces false-positive noise and adds a new rule aligned with the AGENTS.md voice guidelines.

Changes

FernStyles/Current.yml

• Removed old from the time-relative terms list — it was false-positiving on redirect examples (/old-path) and migration docs

.vale.ini

• Added [**/changelog/*.mdx] section that disables FernStyles.Current for changelog entries, since changelogs legitimately use terms like "You can now..." and "latest"

FernStyles/Acronyms.yml

• Added 15 missing acronyms: GTM, PATCH, DELETE, HEAD, OPTIONS, PEM, HMAC, MITM, TOML, ORM, CORS, CIDR, NaN, EOF, CRUD

FernStyles/Headings.yml

• Added Fern product names and common third-party names: Playground, Dashboard, Editor, Writer, Ask Fern, Explorer, Autorelease, Bitbucket, NuGet, Terraform, Gradle, Swagger, Webpack, Vite, Vercel, Netlify, Cloudflare, PostgreSQL, Redis, Stripe, Twilio, and more

FernStyles/Hedges.yml (new)

• New rule that flags hedge words and filler phrases called out in AGENTS.md: simply, just, make sure you, be sure to, you'll want to, please note that, keep in mind, etc.
• Set to suggestion level so it's educational, not blocking

Requested by: Fern Support

[fern-api]

Requested by: devin
Slack thread: View conversation

[github-actions]

🌿 Preview your docs: https://fern-preview-fern-vale-config-analysis.docs.buildwithfern.com/learn