docs: endorse hak-pyth-plugin@0.3.0 and flag the Pyth Core API-key deadline

[jmgomezl]

What

Updates the endorsed version of hak-pyth-plugin to 0.3.0 in both places it is listed, and adds a time-sensitive upgrade note.

Why

1. The two lists were stale and disagreed with each other.

File	Was	Now
docs/PLUGINS.md	0.2.0	0.3.0
packages/core/README.md	0.1.1	0.3.0

2. There is a hard deadline attached to this one.

Pyth Core requires an API key from 2026-08-18 under OP-PIP-100. From that date the Hermes endpoint rejects anonymous price-update requests, so every published version below 0.3.0 stops returning prices. This is not a deprecation warning — the tools start failing.

I verified the enforcement is already live on the upgraded endpoint:

Endpoint	/v2/price_feeds	/v2/updates/price/latest
hermes.pyth.network (current)	200	200
pyth.dourolabs.app/hermes (upgraded)	200	401

0.3.0 adds API-key support and selects the endpoint accordingly: without a key it stays on the current Hermes host (works until 2026-08-18), and setting PYTH_API_KEY switches it to the upgraded authenticated endpoint. Upgrading is non-breaking — existing users keep working, and setting the env var is the whole migration.

Plugin-side change: jmgomezl/hak-pyth-plugin#5.

Scope

Docs only — two version strings and two upgrade notes. No code changes.

Note: both files already fail prettier --check on main today, before this change. I left the existing formatting untouched so the diff stays reviewable rather than reformatting whole files in a docs PR.

[jmgomezl]

Flagging a CI issue I can't resolve from my side: Assignee Check is failing, and it needs a maintainer.

The title-check workflow fails when a PR has no assignee:

- name: Check Assignee
  if: ${{ github.event.pull_request.assignees == null || github.event.pull_request.assignees[0] == null }}
  run: |
    echo "Assignee is not set. Failing the workflow."
    exit 1

I tried to self-assign when opening the PR, but outside contributors don't have the permission:

GraphQL: jmgomezl does not have the correct permissions to execute `ReplaceActorsForAssignable`

Assigning requires write access (or being a repo collaborator), so this check can't pass on a fork-based PR until someone on the team sets an assignee. It looks like it would affect any external contribution, not just this one — worth a look if that isn't intended.

Every other check is green: DCO, Title Check, conventional-pr-title, all five Wiz scanners, and Snyk license + security.

Could a maintainer assign this one so CI can go green? Happy to rebase if anything else is needed.

[jmgomezl]

@MWBlocky @skurzyp-blockydevs — could one of you set an assignee on a few PRs? Sorry to ping directly; this is the one step I can't do from a fork.

Assignee Check is the only failing check on five of my open PRs. Every other check passes on all of them. Outside contributors can't self-assign — GitHub rejects it:

GraphQL: jmgomezl does not have the correct permissions to execute `ReplaceActorsForAssignable`

You assigned #822 and #828 previously, @MWBlocky, which is exactly why those two now go green — so I think this is just the same one-click action, five times.

PR	Files	Status
#1068	2	Endorses hak-pyth-plugin@0.3.0. Time-sensitive — Pyth Core requires an API key from 2026-08-18, and versions below 0.3.0 stop returning prices on that date.
#921	1	Adds hak-uniswap-plugin to the plugin list.
#922	1	Adds hak-ledger-plugin to the plugin list.
#876	1	Adds hak-github-pay-plugin. Rebased today.
#893	9	swap_tokens_tool (core-dex-plugin). Rebased today; Wiz SAST now clean after a validation fix.

All five are MERGEABLE against current main — I rebased the stale ones today and dropped their README.md entries, since the per-plugin list has been consolidated into docs/PLUGINS.md. #822, #823 and #828 are also green and ready.

No rush on the review itself — I know the queue is long. It's only the assignment that's blocking CI.

One thing worth flagging separately: this check appears to fail for any fork-based PR, since no outside contributor can self-assign. If that isn't intended, it may be worth exempting external PRs or having the workflow auto-assign the author. Happy to open a separate issue if that's useful.