#465 docs(strategy): mark the section 6 priorities that have shipped as done

Section 6 of the strategy doc marked only priorities 1 and 5 as done. Priorities 3, 4 and 6 shipped on 2026-09-25 and 2026-09-26 and now carry a Done line with the evidence, in the same format as 1 and 5:

• 3 (memory floor): metrics ledger section 5.45; a 3-file project's index run went from 358 MB to 176 MB peak.
• 4 (diff-seeded impact in explore): #125.
• 6 (delete the TypeScript resolver): #192.

Priority 2 (a persisted resolution origin and confidence on every edge, and declining name-only guesses) is left open: edges carry a provenance column, but I did not find a change that shows both parts done.

README rows checked: none apply (design doc only, no number moved). No CHANGELOG entry.

---

Summary by cubic

Marks section 6 priorities 3, 4, and 6 as done in the strategy doc, and updates the dual-resolver cost into historical context now that #192 removed the TypeScript resolver spine.

• Priority 4's Done line notes the diff-impact task class and first baseline in the agent-eval harness (#460); priority 6's notes it shipped ahead of schedule, ahead of priority 2.
• The "One implementation" cost line now reads as historical: import-resolver.ts remains only as an import-path helper.
• Priority 2 stays open. No README rows or CHANGELOG entries apply.

Summary by CodeRabbit

• Documentation
  • Updated the strategy documentation to record memory profiling, diff-based impact analysis, worktree index seeding, and resolver-port removal as completed.
  • Recorded a peak memory measurement of 176 MB for indexing a three-file project, down from 358 MB, and added the initial evaluation baseline for diff-based impact analysis.
  • Clarified that resolver-port removal happened before the precision work in step 2 was complete, and noted the remaining import-path helper.