fix(audit): exclude .dce-temp partial exports from JSON audit

Operator validation failed when yes_general OOM left truncated exports
under .dce-temp. Audit and archive verification now skip in-progress temps;
smoke covers the partial-temp case. KotOR audit passes with temps present.

docs/plans/2026-06-04-045-fix-audit-exclude-dce-temp-plan.md

@@ -0,0 +1,38 @@
+---
+title: "fix: Exclude .dce-temp from archive JSON audit"
+type: fix
+status: complete
+date: 2026-06-04
+origin: /lfg — KotOR validation audit fails on in-progress partial exports in .dce-temp
+---
+
+# fix: Exclude .dce-temp from archive JSON audit
+
+## Problem
+
+`audit-archive-json.sh` scans every `*.json` under a target output dir. Partial/incomplete exports in `.dce-temp/export.*` (truncated mid-catch-up) fail `jq empty`, so operator validation reports audit failure even when archive files are valid.
+
+Observed: `INVALID .../KotOR_discord_msgs/.dce-temp/export.221726893064454144.kbMFiP/export.json` after yes_general OOM skip preserved partial temp (plan 043).
+
+## Requirements
+
+| ID | Requirement |
+|----|-------------|
+| R1 | `audit-archive-json.sh` skips files under `*/.dce-temp/*` (same as `.dce-meta`) |
+| R2 | `verify-documents-archives.sh` uses consistent exclusion if it scans JSON |
+| R3 | `audit-archive-json-smoke.sh` covers a fixture partial under `.dce-temp` that must not fail audit |
+| R4 | `run-all-smokes.sh` passes (19/19) |
+| R5 | KotOR audit passes while partial temp exists; update `docs/recurring-scrape-merge-readiness.md` |
+
+## Verification
+
+```bash
+./scripts/tests/audit-archive-json-smoke.sh
+DCE_MIN_FREE_MB=0 ./scripts/run-all-smokes.sh
+./scripts/audit-archive-json.sh --config config/scrape-targets.json --target KotOR_discord_msgs
+```
+
+## Out of scope
+
+- Completing yes_general multi-hour catch-up inside LFG
+- Container memory tuning

docs/recurring-scrape-merge-readiness.md

@@ -111,9 +111,13 @@ DCE_MIN_FREE_MB=0 ./scripts/run-operator-validation.sh --sync-gui --per-target -
 | expanded_kotor_discord | pass | pass | validation-resume |
 | eod_discord | pass | pass | validation-resume |
 | DS_Discord_msgs | pass | pass | validation-resume; some channels forbidden |
-| KotOR_discord_msgs | **in progress** | — | plan 044 validation started 2026-06-04 (`logs/kotor-validation-20260604.log`); `yes_general` catch-up + preserve-partial smoke |
+| KotOR_discord_msgs | **scrape pass / audit pass*** | pass* | plan 045: audit excludes `.dce-temp` partials; yes_general catch-up in progress with preserved partial temps (~23–29 MiB) |
 
-**Plan 044 (2026-06-04):** Offline smoke asserts partial temp preserved on OOM skip (channel 134). Host wrapper always writes explicit compose env from `DISCORD_TOKEN_FILE` (fixes auth-retry when shell exports a stale `DISCORD_TOKEN`). `run-all-smokes.sh` → 19/19 pass.
+\* Audit failed before plan 045 because truncated partial exports under `.dce-temp/` were scanned as archives. After fix, audit passes while partial temps exist.
+
+**Plan 045 (2026-06-04):** `audit-archive-json.sh` and `verify-documents-archives.sh` skip `*/.dce-temp/*` (in-progress partial exports). Salvage run 2026-06-03: 7 merged, 17 unchanged, 3 skipped (+5404 messages); yes_general OOM-skipped with partial temps preserved for next salvage.
+
+**Plan 044 (2026-06-04):** Offline smoke asserts partial temp preserved on OOM skip (channel 134). Host wrapper prefers `DISCORD_TOKEN_FILE` over inherited shell tokens. `run-all-smokes.sh` → 19/19 pass.
 
 **KotOR / yes_general (plan 040–043):** Incremental `--after` works for all channels; most return `UNCHANGED` in seconds. `yes_general` archive last message was **2021-01-17** — the first catch-up legitimately fetches years of history. Prior bug: OOM skip **deleted** partial temp exports, causing re-download loops. Plan 043 preserves partial temps and salvages on next run.
 

scripts/audit-archive-json.sh

@@ -35,7 +35,7 @@ audit_dir() {
     fi
     printf 'INVALID\t%s\n' "$file_path"
     FAILURES=$((FAILURES + 1))
-  done < <(find "$output_dir" -type f -name '*.json' ! -path '*/.dce-meta/*' -print0 2>/dev/null)
+  done < <(find "$output_dir" -type f -name '*.json' ! -path '*/.dce-meta/*' ! -path '*/.dce-temp/*' -print0 2>/dev/null)
 }
 
 main() {

scripts/tests/audit-archive-json-smoke.sh

@@ -21,6 +21,9 @@ JSON
 
 printf '{"messages":[\n' >"$ARCHIVE_ROOT/bad/truncated [222].json"
 
+mkdir -p "$ARCHIVE_ROOT/good/.dce-temp/export.111.PARTIAL"
+printf '{"messages":[\n' >"$ARCHIVE_ROOT/good/.dce-temp/export.111.PARTIAL/export.json"
+
 cat >"$CONFIG_PATH" <<JSON
 {
   "archive_root": "$ARCHIVE_ROOT",

scripts/verify-documents-archives.sh

@@ -28,7 +28,7 @@ require_command() {
 
 count_archive_json() {
   local output_dir=$1
-  find "$output_dir" -type f -name '*.json' ! -path '*/.dce-meta/*' 2>/dev/null | wc -l | tr -d ' '
+  find "$output_dir" -type f -name '*.json' ! -path '*/.dce-meta/*' ! -path '*/.dce-temp/*' 2>/dev/null | wc -l | tr -d ' '
 }
 
 count_seeded_channel_ids() {
@@ -42,7 +42,7 @@ count_seeded_channel_ids() {
     if [[ "$file_name" =~ \[([0-9]{16,22})\]\.json$ ]]; then
       printf '%s\n' "${BASH_REMATCH[1]}"
     fi
-  done < <(find "$output_dir" -type f -name '*.json' ! -path '*/.dce-meta/*' -print0 2>/dev/null) | sort -u | wc -l | tr -d ' '
+  done < <(find "$output_dir" -type f -name '*.json' ! -path '*/.dce-meta/*' ! -path '*/.dce-temp/*' -print0 2>/dev/null) | sort -u | wc -l | tr -d ' '
 }
 
 count_channel_map_entries() {