feat(signals): implement issue quality reports for opportunity ranking

[JSONbored]

Problem

Gittensory has repo queue data, but it does not yet deeply classify which issues are actually useful to pursue. Without issue quality, miners can still waste time on stale, vague, duplicate-prone, or already-solved work.

Desired behavior

Populate issue quality reports and feed them into MCP opportunities, local preflight, and agent next-action planning.

Implementation notes

• Use existing cached GitHub issue data and repo lane context.
• Classify issues as actionable, too broad, stale, duplicate-prone, already solved, needs maintainer context, or good candidate.
• Treat direct-PR-only and issue-discovery repos differently.
• Keep classification deterministic in v1.

Public/private boundaries

Issue quality can be used in private MCP/API recommendations. Public GitHub output should only include sanitized, actionable context and should not shame authors or maintainers.

Acceptance criteria

• MCP can explain why an issue is or is not worth pursuing.
• Direct-PR-only repos do not encourage noisy issue filing.
• Issue-discovery repos can surface good candidate issues.
• Duplicate/WIP risk affects opportunity ranking.
• Large repos stay within Worker budget.

Tests

• Stale issue fixture.
• Duplicate-prone issue fixture.
• Already-solved issue fixture.
• Vague issue fixture.
• High-quality issue fixture.
• Direct-PR-only lane fixture.
• Issue-discovery lane fixture.