chore: dispatch ci-timing.yml to re-baseline after #2258 and #2244, and attach the report to #2254 #2554

From the 2026-11-01 conventions audit, re-verified on 73add05a3. process.rst, "CI Timing Review": after a merged ci PR changes a job's shape or selection, dispatch ci-timing.yml so the next report compares like with like; a HARD failure's issue carries the ci-timing-report artifact and the slowest-module table.

• ci-timing.yml has only scheduled runs, so its last complete baseline (run 40288101249) predates both changes: gh api 'repos/JarryShaw/PyPCAPKit/actions/workflows/ci-timing.yml/runs?event=workflow_dispatch' -q .total_count prints 0.
• ci(ordering): shard the project ordering leg and cache the tier guard's workflow parse (#2254) #2258 split unittest-ordering's project cell into project (half 1) and project (half 2); ci: let docs-only and test-only pull requests skip the heavy matrices (#2243) #2244 let docs-only and test-only PRs skip the heavy matrices. Both carry ci, merged after that run, and neither thread mentions a dispatch. Each run's report.json is the next run's baseline, so the next scheduled run compares against the unsplit Plain unittest ordering (project).
• ci(timing): the project ordering leg will hit its 20-minute step cap within about a week #2254, the HARD-failure issue for run 40288101249, names the run in prose but attaches neither the artifact nor the table (gh issue view 2254 --json body,comments). The artifact is still downloadable: gh api repos/JarryShaw/PyPCAPKit/actions/runs/40288101249/artifacts shows expired: false.

Fix: gh workflow run ci-timing.yml and confirm it writes a report; then comment the artifact link and the slowest-module table from its report.md on #2254 (closed; the comment still belongs there).