docs(performance): retain post-import gathering and upgraded runtime observations

performance/projects/installed-observation/README.md

@@ -70,3 +70,37 @@ imports, reading complete relevant passages, project attribution, contradictory
 proposal review remain separate work. Agent reading/reasoning, manual tool calls, remote imports,
 model accuracy and token costs were not timed. The repeated phase returns evidence again; it does
 not establish an incremental AI-analysis cache.
+
+## Follow-up: import coverage and the newer runtime
+
+A native selected-folder import finished outside the timed phases, and selected missing messages
+were retrieved. A new observation reused the request on Memo 0.79.0/Core 0.367.0, then another used
+published Memo 0.81.0/Core 0.370.0. The archive and index remain mutable. The three phases in each
+observation returned the same direct reference and excerpt-hash sets.
+
+| State | Direct / combined refs | Calls per phase | Pagination | No-progress stop |
+| --- | ---: | ---: | --- | --- |
+| Earlier 0.79 observation above | 67 / 144 | 1 | Finished | No |
+| After source import, 0.79 | 92 / 92 | 3 | Incomplete | Yes |
+| After source import, 0.81 | 93 / 170 | 1 | Finished | No |
+
+The 0.79 follow-up adds 25 documents and retains all previous direct references, but one search
+unit remains partial with a validation error. Empty continuations stop after unchanged progress.
+An initial attempt failed before writing a completed measurement and is retained privately too.
+The 0.81 observation retains all 92 direct references and their excerpts, adding one document;
+no unit stalls. Source imports are separate from pagination, and a larger direct count alone does
+not establish better surrounding context or answer quality.
+
+The 0.81 phases took 1,020.28 / 1,063.35 / 1,131.05 ms, including their single CLI invocation.
+Those timings are observations, not a controlled version comparison or end-to-end workflow cost.
+The registry package matches its validated release artifact (SHA-256
+`e88a1b8290435b37e2175d5d1931980bf0d77dab26a2336dfa996cccbb08e8d1`), and packaged/installed
+CLI and stdio MCP smoke checks pass on invented data with temporary state. Initial registry tarball
+404s resolved before the registry bytes were compared. Runtime fingerprints match within each trial.
+
+[0.79 after import](post-import-079-totals.json) and [0.81 after import](post-import-081-totals.json)
+are produced by the unchanged aggregate projector above. Each includes its private measurement
+checksum, three phase totals and qualifications. Neither includes the project's name, query,
+source references, source text, per-question answers or labels. Original totals remain unchanged.
+Finished pagination still leaves remote completeness and semantic interpretation unknown; source
+review and proposed-finding decisions are separate from this measurement.

performance/projects/installed-observation/post-import-079-totals.json

@@ -0,0 +1,78 @@
+{
+  "memo": "0.79.0",
+  "core": "0.367.0",
+  "base": "0.47.1",
+  "node": "v24.19.0",
+  "private_measurement_sha256": "9229088fecca9d5095e92267391148e9969de4d1795bea494c3b9ff913921cf9",
+  "phases": [
+    {
+      "phase": "index_as_stored",
+      "invocations": 3,
+      "wall_ms": 2334.37,
+      "direct_references": 92,
+      "surrounding_references": 48,
+      "combined_references": 92,
+      "direct_by_kind": {
+        "message": 41,
+        "document": 50,
+        "meeting": 1
+      },
+      "direct_excerpt_bytes": 34099,
+      "distinct_surrounding_excerpt_bytes": 36942,
+      "distinct_surrounding_excerpts": 187,
+      "charged_read_budget_bytes": 2064216,
+      "charged_surrounding_budget_bytes": 36942,
+      "pagination_finished": false,
+      "no_progress_stop": true,
+      "oversized_references": 0,
+      "resolved_sources": 14
+    },
+    {
+      "phase": "default_catch_up",
+      "invocations": 3,
+      "wall_ms": 2655.98,
+      "direct_references": 92,
+      "surrounding_references": 48,
+      "combined_references": 92,
+      "direct_by_kind": {
+        "message": 41,
+        "document": 50,
+        "meeting": 1
+      },
+      "direct_excerpt_bytes": 34099,
+      "distinct_surrounding_excerpt_bytes": 36942,
+      "distinct_surrounding_excerpts": 187,
+      "charged_read_budget_bytes": 2064216,
+      "charged_surrounding_budget_bytes": 36942,
+      "pagination_finished": false,
+      "no_progress_stop": true,
+      "oversized_references": 0,
+      "resolved_sources": 14
+    },
+    {
+      "phase": "default_repeat",
+      "invocations": 3,
+      "wall_ms": 2435.86,
+      "direct_references": 92,
+      "surrounding_references": 48,
+      "combined_references": 92,
+      "direct_by_kind": {
+        "message": 41,
+        "document": 50,
+        "meeting": 1
+      },
+      "direct_excerpt_bytes": 34099,
+      "distinct_surrounding_excerpt_bytes": 36942,
+      "distinct_surrounding_excerpts": 187,
+      "charged_read_budget_bytes": 2064216,
+      "charged_surrounding_budget_bytes": 36942,
+      "pagination_finished": false,
+      "no_progress_stop": true,
+      "oversized_references": 0,
+      "resolved_sources": 14
+    }
+  ],
+  "app_model_calls": 0,
+  "model_tokens": null,
+  "qualified_as": "Mutable local archive observation; not complete remote coverage, semantic recall, model quality, agent time, actual I/O or cold-cache speedup."
+}

performance/projects/installed-observation/post-import-081-totals.json

@@ -0,0 +1,78 @@
+{
+  "memo": "0.81.0",
+  "core": "0.370.0",
+  "base": "0.47.1",
+  "node": "v24.19.0",
+  "private_measurement_sha256": "906121f3954b0004616d36c58597347aa750b446cfa5d1d67a06ed1d399504e5",
+  "phases": [
+    {
+      "phase": "index_as_stored",
+      "invocations": 1,
+      "wall_ms": 1020.28,
+      "direct_references": 93,
+      "surrounding_references": 144,
+      "combined_references": 170,
+      "direct_by_kind": {
+        "message": 41,
+        "document": 51,
+        "meeting": 1
+      },
+      "direct_excerpt_bytes": 34604,
+      "distinct_surrounding_excerpt_bytes": 49604,
+      "distinct_surrounding_excerpts": 287,
+      "charged_read_budget_bytes": 1898846,
+      "charged_surrounding_budget_bytes": 55867,
+      "pagination_finished": true,
+      "no_progress_stop": false,
+      "oversized_references": 0,
+      "resolved_sources": 14
+    },
+    {
+      "phase": "default_catch_up",
+      "invocations": 1,
+      "wall_ms": 1063.35,
+      "direct_references": 93,
+      "surrounding_references": 144,
+      "combined_references": 170,
+      "direct_by_kind": {
+        "message": 41,
+        "document": 51,
+        "meeting": 1
+      },
+      "direct_excerpt_bytes": 34604,
+      "distinct_surrounding_excerpt_bytes": 49604,
+      "distinct_surrounding_excerpts": 287,
+      "charged_read_budget_bytes": 1898846,
+      "charged_surrounding_budget_bytes": 55867,
+      "pagination_finished": true,
+      "no_progress_stop": false,
+      "oversized_references": 0,
+      "resolved_sources": 14
+    },
+    {
+      "phase": "default_repeat",
+      "invocations": 1,
+      "wall_ms": 1131.05,
+      "direct_references": 93,
+      "surrounding_references": 144,
+      "combined_references": 170,
+      "direct_by_kind": {
+        "message": 41,
+        "document": 51,
+        "meeting": 1
+      },
+      "direct_excerpt_bytes": 34604,
+      "distinct_surrounding_excerpt_bytes": 49604,
+      "distinct_surrounding_excerpts": 287,
+      "charged_read_budget_bytes": 1898846,
+      "charged_surrounding_budget_bytes": 55867,
+      "pagination_finished": true,
+      "no_progress_stop": false,
+      "oversized_references": 0,
+      "resolved_sources": 14
+    }
+  ],
+  "app_model_calls": 0,
+  "model_tokens": null,
+  "qualified_as": "Mutable local archive observation; not complete remote coverage, semantic recall, model quality, agent time, actual I/O or cold-cache speedup."
+}