Plan and track expansion of destinations.json from 11 to 100+ entries with balanced label distribution

[knytcomics-ui]

Problem

11 destinations is enough to prove the pipeline runs end-to-end, but far too small a sample for grydlock-research's accuracy/false-positive measurements (per the README's stated goal) to be statistically meaningful. This issue is the tracking/coordination point for the fixture-volume work implied by the many entity-type and transaction-shape expansion issues already in this batch, so growth happens against a plan rather than ad hoc.

Requirements

• Define a target composition (e.g. propose ~40% clean / 30% suspicious / 30% malicious at 100+ total entries, informed by whatever real-world base rate grydlock-research is trying to approximate — coordinate with that repo's maintainers if the target base rate is documented there).
• Ensure the target composition explicitly allocates coverage across every entity type and risk pattern introduced by the other fixture issues in this batch (accounts, assets, contracts, muxed accounts, liquidity pools, sponsored/signer-takeover/memo-based/cold-start patterns) rather than just duplicating more of the existing 11 patterns.
• Track progress toward the target as a checklist (in this issue, the README roadmap, or a project board — pick one canonical location and say so) so partial progress across many separate fixture PRs stays visible.

Acceptance Criteria

• A concrete target composition (counts per label, checklist of entity/pattern types) is written down and linked from the README roadmap.
• This issue serves as (or points to) the single tracking location for the expansion effort across all related PRs.

Part of the Gryd-lock organization. This repo supplies the labelled testnet fixtures and stub scores consumed by grydlock-oracle-adapter's StubOracle and evaluated end-to-end by grydlock-research.