[15] Add per-contract CI jobs and a coverage job to detect regressions faster #15

Problem Statement

.github/workflows/ci.yml defines a single CI job that runs all six contracts in sequence. There is no per-contract parallelism, no coverage measurement, and no separate jobs for fmt, clippy, test, build. A regression in any one contract blocks feedback for all.

Why It Matters

• Slow CI is a contributor-experience killer. Per-contract jobs halve or third the critical path.
• No coverage = no idea what the tests do or don't cover.

Expected Outcome

• Split into:
  • contracts-fmt: runs across all crates in parallel.
  • contracts-clippy: same.
  • contracts-test-{reward-pool,course-registry,badge-nft,quest-engine,stake-vault,governance} matrix.
  • contracts-build-wasm: builds WASM artifacts and uploads.
  • contracts-coverage: uses cargo-tarpaulin or llvm-cov to generate a coverage report, posted back to PRs as a comment.
• A reusable composite action .github/actions/setup-stellar for Stellar CLI.

Acceptance Criteria

• CI runtime reduced by ≥30% via parallelization.
• Coverage threshold >80% enforced (or measured and reported).
• Per-contract test results clearly indicate which crate failed.

Implementation Notes

• Use cargo-nextest for per-test parallelism.
• Cache target/ per crate with Swatinem/rust-cache already in workflow.

Files / Modules Affected

• .github/workflows/ci.yml (rewrite).
• New .github/actions/setup-stellar/action.yml.

Dependencies

None.

Difficulty

Medium.

Estimated Effort

1–2 days.

Suggested Labels

ci, devops, P2