diff --git a/ROADMAP.md b/ROADMAP.md
index 5cb60288..017a1318 100644
--- a/ROADMAP.md
+++ b/ROADMAP.md
@@ -7555,3 +7555,5 @@ Original filing (2026-04-18): the session emitted `SessionStart hook (completed)
 ||||||| parent of 52161c78 (docs(roadmap): add #330 — resume mode stats/cost always zero)
 
 330. **`/stats` and `/cost` in `--resume` mode always return zero token counts regardless of saved session usage** — dogfooded 2026-04-29 by Jobdori on current main (`e7074f4`). Running `claw --output-format json --resume latest /stats` and `claw --output-format json --resume latest /cost` both return `{"input_tokens": 0, "output_tokens": 0, "total_tokens": 0, ...}` even when the session file was created by a real interactive claw run. The same zero-fill is produced by `claw --output-format json status` (usage sub-object). Inspection of the default session path (`~/.claw/sessions/.../session-*.jsonl`) shows the file has only 2 events (`session_meta`, `message`) with no usage/token records embedded. Two candidate root causes: (a) session serialization never writes usage events into the JSONL file even after real prompt exchanges, so resume-mode has no usage to replay; (b) resume-mode stat accumulation reads usage from memory-side counters that are always zero because no API calls happened in the resume-only invocation. Either way, the operator effect is the same: `--resume` + `/stats`/`/cost` cannot report what the session actually consumed. **Required fix shape:** (a) persist per-turn usage records (input/output/cache tokens per exchange) into the session JSONL at write time so resume-mode can reconstruct cumulative counts by replay; (b) expose a `"total_usage"` summary block at the tail of every session JSONL so resume-mode can read it in O(1) without full replay; (c) ensure `/stats` and `/cost` in resume-mode sum the persisted usage, not memory-side live counters; (d) add regression coverage proving `--resume SESSION.jsonl /stats` returns non-zero token counts after a session that made real API calls. **Why this matters:** token usage visibility is critical for quota management and cost attribution; if `--resume` mode always shows zeros, operators and orchestration lanes cannot trust usage data from saved sessions and must rely on external billing dashboards instead of in-band tooling. Source: Jobdori live dogfood on mengmotaHost, claw-code `e7074f4`, 2026-04-29.
+
+696. **`claw compact` (and likely other REPL-only commands) hangs indefinitely in non-interactive mode with no TTY — there is no timeout, no stdin-closed guard, and no `--output-format json` fast-exit path** — dogfooded 2026-05-25 on `bb2a9238`. Running `./rust/target/debug/claw compact --output-format json --help </dev/null` (or `compact` without `--help`) never returns: the process blocks waiting for interactive input that will never arrive. No timeout fires, no error is emitted on stdout, no `{"kind":"error",...}` JSON is produced. The caller must `kill` the process externally. This is a clawability gap: any orchestrator or claw that probes `compact` to discover its schema or trigger non-interactive compaction will hang until a watchdog kills it. **Required fix shape:** (a) detect stdin EOF / non-TTY at startup for REPL-only commands and emit `{"kind":"error","error_kind":"interactive_only","message":"claw compact requires an interactive terminal"}` with exit 1 instead of blocking; (b) honour `--output-format json` as an implicit non-interactive signal — if the flag is set and no TTY is present, exit immediately with a typed error; (c) add a global `--timeout <seconds>` flag or default 30 s watchdog for non-interactive invocations; (d) add regression coverage proving `claw compact </dev/null` exits non-zero within 1 s with a structured error. **Why this matters:** automation layers probe commands via `</dev/null` to discover schema; a hung process blocks the probe indefinitely and requires external timeout management, making compact undiscoverable and unscriptable. Source: Jobdori dogfood on `bb2a9238`, 2026-05-25.