feat: near-miss deep analysis — per-task failure details

[jpleva91]

Problem

Near-misses (50%+ tests passing) are the highest-leverage improvement targets, but the sentinel report only shows pass/fail counts. Need detailed per-assertion analysis for near-miss tasks.

Acceptance Criteria

• For each near-miss task, extract individual test names and their pass/fail status from ctrf.json
• Report which specific assertions failed (not just count)
• Add "Near-Miss Deep Dive" section to sentinel markdown report

Hints

• ctrf.json at task_dir/verifier/ctrf.json has per-test results
• Near-miss tasks identified by classify_task() returning "near_miss"

---

Generated by /forge cascade

[jpleva91]

@claude Please implement this issue. Read the full codebase first to understand the architecture, then implement per the acceptance criteria. Run go test ./... and go build ./... to verify.