Add GitHub-auditable official results and trace export

Author: sjarmak

README.md

@@ -238,6 +238,31 @@ The report generator produces:
 
 See `python3 scripts/generate_eval_report.py --help` for all options.
 
+### Publishable Official Results + Trace Browser
+
+To export GitHub-friendly official results (valid scored tasks only) with parsed
+trace summaries and local browsing UI:
+
+```bash
+python3 scripts/export_official_results.py \
+  --runs-dir ./runs/official/ \
+  --output-dir ./docs/official_results/
+```
+
+This writes:
+- `docs/official_results/README.md` -- run/config score summary
+- `docs/official_results/runs/*.md` -- per-run task tables
+- `docs/official_results/tasks/*.md` -- per-task metrics + parsed tool/trace view
+- `docs/official_results/data/official_results.json` -- machine-readable dataset
+- `docs/official_results/audits/*.json` -- per-task audit artifacts (checksums + parsed trace events)
+- `docs/official_results/index.html` -- interactive local browser
+
+Serve locally:
+
+```bash
+python3 scripts/export_official_results.py --serve
+```
+
 For the full multi-layer evaluation pipeline (verifier, LLM judge, statistical analysis, dual-score reporting), see [docs/EVALUATION_PIPELINE.md](docs/EVALUATION_PIPELINE.md).
 
 ---

docs/OFFICIAL_RESULTS_BROWSER.md

@@ -0,0 +1,42 @@
+# Official Results Browser
+
+Use this workflow to publish valid official scores with easy-to-view parsed traces.
+
+## What It Exports
+
+`python3 scripts/export_official_results.py` scans `runs/official/` and exports only valid scored tasks (status `passed`/`failed` with numeric reward) into a static bundle:
+
+- `docs/official_results/README.md` - run/config score summary
+- `docs/official_results/runs/*.md` - per-run task tables
+- `docs/official_results/tasks/*.md` - per-task metrics and parsed trace/tool summaries
+- `docs/official_results/data/official_results.json` - machine-readable data
+- `docs/official_results/audits/*.json` - per-task audit payloads with trace parsing and SHA256 checksums
+- `docs/official_results/index.html` - local interactive browser
+
+## Usage
+
+```bash
+python3 scripts/export_official_results.py \
+  --runs-dir ./runs/official/ \
+  --output-dir ./docs/official_results/
+```
+
+Filter to specific run(s):
+
+```bash
+python3 scripts/export_official_results.py \
+  --run ccb_mcp_compliance_haiku_20260226_205845 \
+  --run ccb_mcp_domain_haiku_20260226_205845
+```
+
+Serve locally after export:
+
+```bash
+python3 scripts/export_official_results.py --serve
+```
+
+## Notes
+
+- The exporter prefers `task_metrics.json` when present and falls back to transcript parsing for tool-call extraction.
+- Task pages link to bundled `audits/*.json` so GitHub viewers can audit without local `runs/official/`.
+- If `runs/official/MANIFEST.json` exists, export is automatically scoped to run directories tracked in the manifest.