# Phase 25.4 — Smoke Probe Report **Purpose:** end-to-end verification that the Phase 25.4 nested-migration writers no longer create flat psyche/worker perch dirs as side effects. **Method:** before-and-after comparison of `.planning/quick/260523-paw-perch-audit-psyche-worker-flat-vs-nested/probe.ps1` output against the operator's `%LOCALAPPDATA%\spt\owlery` filesystem. **Source:** CONTEXT.md D-01 + D-02 + D-05 + D-07; RESEARCH.md §3 (pre-deploy evidence). --- ## Pre-Deploy Probe (2026-05-23 18:21 PT — captured at RESEARCH time) Source: RESEARCH.md §3 (Section 3 — On-Disk Evidence). Active live agent at capture time: `todlando`. ### todlando-psyche (FLAT) — every file fresh, active writer ``` DIR: …\owlery\todlando-psyche dir-mtime=2026-05-23 18:18:13 2026-05-22 16:54 inbox (created at first boot; rarely re-touched) 2026-05-23 17:18 0 .has-messages (FRESH, ~1h) 2026-05-23 18:11 0 .idle-ready (FRESH, ~10min) 2026-05-22 16:55 0 .more-done (stale) 2026-05-23 18:11 315 info.json (FRESH — written by poll.rs every iter) 2026-05-23 17:18 0 ready (FRESH) 2026-05-23 17:18 36864 spool.db (FRESH — WAL active) 2026-05-23 18:11 44 status (FRESH — poll.rs status line) 2026-05-23 17:18 139 wrapper-state.json (FRESH — claude.rs:201 publish) ``` ### todlando/nested/todlando-psyche (NESTED) — fixtures only, no live writer ``` DIR: …\owlery\todlando\nested\todlando-psyche dir-mtime=2026-05-23 17:18:21 2026-04-19 04:06 inbox (one month old) 2026-05-22 03:28 316 info.json (one day old — start.rs nested initial write) 2026-05-23 17:18 0 ready (~1h — refreshed by start.rs at last boot? See below) 2026-05-21 17:54 12288 spool.db (two days old) 2026-05-22 03:15 44 status (one day old) 2026-05-22 01:55 114 wrapper-state.json (one day old) ``` The `ready` file at the nested path was refreshed at 17:18 — coinciding with the LAST `$LIVE start` (or the post-handoff publish via `start.rs:524 / 834`). The wrapper's loop-exit predicate at `wrapper/mod.rs:1168` is reading THAT file. Nothing in the steady-state poll loop refreshes the nested ready file. **Survives only because no code path deletes it.** ### Workers — flat dominant, nested fragmentary Flat (7 dirs, last hour activity): ``` todlando-w81/ .owl-aware (17:28) spool.db (17:28) todlando-w82/ .owl-aware (17:29) spool.db (17:29) todlando-w83/ .owl-aware (17:48) spool.db (17:48) todlando-w84/ .owl-aware (17:52) spool.db (17:52) todlando-w85/ .owl-aware (17:57) spool.db (17:57) todlando-w86/ .owl-aware (17:59) spool.db (17:59) todlando-w87/ .owl-aware (18:04) spool.db (18:04) ``` These flat dirs have NO `info.json` and NO `ready` file — `enumerate_perches::push_if_perch` skips them silently, so they don't appear in `live list`. They are pure side-effect of hook_check.rs:141 (`.owl-aware`) + spool.rs:12-18 (`spool.db` autocreate). Nested (2 dirs, w85 and w87 only): ``` todlando/nested/todlando-w85/ info.json (17:57) ready (17:57) inbox/ — fully formed nested perch todlando/nested/todlando-w87/ info.json (18:04) ready (18:04) inbox/ — fully formed nested perch ``` w81-w84 and w86 have NO nested counterpart. So `spawn_worker_perch` only fired (or only succeeded) for w85 and w87 — the other 5 workers got their flat dirs from spool/inbox auto-create paths WITHOUT a properly registered perch ever existing. Implication: those subagents never got `spawn_worker_perch` called for them (or it failed silently). **Worth investigating as a separate issue** — Open Question 1. ### Other agents (offline at time of research) - `doyle` (offline, last started 5/22): has `nested/doyle-psyche/` with matching flat counterpart from before the most recent migration. No active writer. - `webber` (offline): nested-only `nested/webber-psyche/`. **Pure nested case** — confirms the migration DID land for some agents at some point; current writes lost the nested-aware shape. - `dunsen` (offline, last started 5/23 17:17): flat `dunsen/` Self-level only. No psyche dir at either layer (psyche was cleaned up by the soft stop). --- ## Post-Deploy Probe (2026-05-23 22:01:54 PT — captured after Phase 25.4 v1.11.10 deploy + `$LIVE start probetest` + subagent spawn) Probe output captured to `post-probe.txt`. Reproduced inline: ``` ROOT: C:\Users\decid\AppData\Local\spt\owlery === TOP LEVEL === Name LastWriteTime ---- ------------- deployah 5/23/2026 7:39:30 PM doyle 5/23/2026 7:39:30 PM dunsen 5/23/2026 10:01:02 PM executor 5/23/2026 7:39:30 PM mica 5/23/2026 7:39:30 PM probetest 5/23/2026 10:01:54 PM todlando 5/23/2026 7:39:30 PM webber 5/23/2026 10:01:02 PM witty 5/23/2026 7:39:30 PM .registry 5/23/2026 9:53:58 PM .session-cache 5/23/2026 10:00:52 PM === FRESH FILES (last 2h) under owlery (recursive) === 0.8 min C:\Users\decid\AppData\Local\spt\owlery\probetest\.idle-ready 1 min C:\Users\decid\AppData\Local\spt\owlery\probetest\spool.db 1 min C:\Users\decid\AppData\Local\spt\owlery\.session-cache 1.2 min C:\Users\decid\AppData\Local\spt\owlery\probetest\.working-counter 7.9 min C:\Users\decid\AppData\Local\spt\owlery\.registry 7.9 min C:\Users\decid\AppData\Local\spt\owlery\probetest\status 7.9 min C:\Users\decid\AppData\Local\spt\owlery\probetest\info.json 7.9 min C:\Users\decid\AppData\Local\spt\owlery\probetest\nested\probetest-psyche\.idle-ready 7.9 min C:\Users\decid\AppData\Local\spt\owlery\probetest\nested\probetest-psyche\status 7.9 min C:\Users\decid\AppData\Local\spt\owlery\probetest\nested\probetest-psyche\info.json 7.9 min C:\Users\decid\AppData\Local\spt\owlery\probetest\nested\probetest-psyche\spool.db 7.9 min C:\Users\decid\AppData\Local\spt\owlery\probetest\nested\probetest-psyche\wrapper-state.json 8.1 min C:\Users\decid\AppData\Local\spt\owlery\probetest\nested\probetest-psyche\ready 8.1 min C:\Users\decid\AppData\Local\spt\owlery\probetest\ready 10.4 min C:\Users\decid\AppData\Local\spt\owlery\dunsen\.idle-ready 89.4 min C:\Users\decid\AppData\Local\spt\owlery\dunsen\info.json === FLAT PSYCHE DIRS === === FLAT WORKER DIRS (-w[N]) === === NESTED CHILD DIRS (owlery/*/nested/*/) === DIR: C:\Users\decid\AppData\Local\spt\owlery\probetest\nested\probetest-psyche dir-mtime=05/23/2026 22:01:54 5/23/2026 9:53:49 PM inbox 5/23/2026 9:53:58 PM 0 .idle-ready 5/23/2026 9:53:58 PM 181 info.json 5/23/2026 9:53:49 PM 0 ready 5/23/2026 9:53:58 PM 12288 spool.db 5/23/2026 9:53:58 PM 45 status 5/23/2026 9:53:58 PM 138 wrapper-state.json (+ doyle/nested/doyle-psyche, todlando/nested/todlando-psyche, todlando/nested/todlando-w85, webber/nested/webber-psyche — all pre-existing per CONTEXT D-05; old mtimes confirm no fresh flat-writer side effects) ``` --- ## Diff Annotations ### New test agent (`probetest`) — expectations - [x] `probetest-psyche/` flat dir does NOT exist — confirmed empty `=== FLAT PSYCHE DIRS ===` - [x] `probetest/nested/probetest-psyche/` EXISTS with fresh files (info.json, ready, status, inbox/, spool.db, .idle-ready, wrapper-state.json — all mtimes ≤8.1 min) - [x] `probetest-wN/` flat worker dirs do NOT exist — confirmed empty `=== FLAT WORKER DIRS ===` - [~] `probetest/nested/probetest-wN/` not observed for spawned subagent — but ALSO not observed flat → migrated writers correctly produce zero flat dirs; spawn_worker_perch evidently did not fire a fully-formed nested perch for the one-shot subagent. Carry over RESEARCH Open Q1 (worker-perch asymmetry) as separate investigation — independent of Phase 25.4 GOAL (no flat side-effect dirs). ### RESEARCH-era orphans (expected to persist per CONTEXT D-05) - [x] No fresh files inside `todlando-psyche/` or any `todlando-w*` — all such dirs cleaned up from top-level listing (only `todlando/` Self dir remains; nested entries have mtimes ≥1 day old) - [x] `doyle/nested/doyle-psyche` and `webber/nested/webber-psyche` show no fresh writes — old mtimes intact ### Worker-mystery (RESEARCH Open Q1) resolution - [ ] Worker-perch asymmetry persists at low signal — one-shot subagent did not create either flat OR nested perch dir. File follow-up under `.planning/debug/-worker-perch-asymmetry-persists.md` per CONTEXT D-07. Out of Phase 25.4 GOAL scope: phase asserts NO flat side effects, which holds. Whether spawn_worker_perch fires reliably is a separate registration question. ### Follow-up fixes applied during probe (in-scope finishing) - [x] **`.more-done` gate predicate** (`src/owl/hook_idle.rs:137`) — flipped from flat `owlery::ready_file(psyche_id)` to nested `perch_path::resolve_ready_file(psyche_id, Explicit(owl_id))`. Pre-fix: gate always failed post-25.4 (flat ready file never exists) → sentinel never written → wrapper echo-commune fires silently dead. Verified via reprobe: `.more-done` sentinel landed at `probetest/nested/probetest-psyche/.more-done` after Stop hook fired on v1.11.11 binary. Commit `48b61be`. - [x] **Stop hook version-change early-return** (`src/owl/hook_idle.rs::run`) — dropped `VersionPrompt::Emitted => return` short-circuit. Pre-fix: any Stop hook firing a version-change emission skipped both `set_idle_ready` and `spawn_echo_commune_if_live` (both pure filesystem sentinel writers, orthogonal to spool/stdout) → `.more-done` dropped for one fire per version transition. Commit `72b4263`. Verified by reasoning; operator accepted without re-proof (sentinel already present for current session from prior fire under v1.11.11). ### Why was this missed by the Phase 25.4 planner? Plan 25.4-04 listed `src/owl/hook_idle.rs` in `files_modified` and migrated the `.more-done` WRITER at line 153. The READER predicate at line 137 — `owlery::ready_file(psyche_id).exists()` — went unmigrated because: 1. **Reader audit was call-site scoped, not file-scoped.** RESEARCH Class-E enumerated direct path-composition reader calls (`info_file`, `perch_dir`, `ready_file`). It captured discrete sites but did not require: "for every file we touch the writer in, sweep all sibling readers in the same function." A gate predicate is a degenerate reader (boolean `.exists()` on a composed path) that the audit's call-site index treated as low-risk. 2. **Plan 04 vs Plan 06 split routed the reader to the wrong plan.** Plan 04 = lower-layer writer migration. Plan 06 = eight Class-E reader sites. The gate predicate sat inside Plan 04's file but was conceptually Plan 06 work. Neither plan claimed it. It fell between the seams. 3. **Tests pinned the broken contract.** The pre-fix test helper `touch_psyche_ready` wrote the ready sentinel at the FLAT path — exactly matching the flat-gate. The test setup secretly co-evolved with the bug, so all 7 tests passed against a gate that would never fire in production. 4. **No integration test exercised the .more-done writer end-to-end.** The integration suite added in Plan 25.4-07 (`perch_path_no_flat_psyche.rs`, `perch_path_no_flat_worker.rs`) asserts the NEGATIVE invariant (no flat dirs created). It does not assert the POSITIVE invariant (`.more-done` actually fires when conditions are met). A POSITIVE-side smoke would have caught this in CI. **Mitigation for next phase of this kind:** when migrating a writer, grep the SAME file for `owlery::*_file(`, `owlery::perch_dir(`, `owlery::info_file(` predicates and gate calls — not just the writer call. Add to PATTERNS.md if recurrence likely. The Stop hook version-change early-return bug (commit `72b4263`) was unrelated to Phase 25.4 — pure Phase 34 over-broad ordering invariant that happened to surface during 25.4 reprobe. --- ## Resolution **Status:** PASS (with 2 in-scope finishing fixes applied during checkpoint loop) **Notes:** - Phase 25.4 GOAL — "psyche/worker writers no longer create flat perch dirs as side effects" — confirmed by probe: zero flat psyche dirs, zero flat worker dirs, all probetest writes nested-only. - Two seam bugs caught and fixed in-checkpoint: `.more-done` gate predicate (Phase 25.4 miss) + Stop hook version-change early-return (Phase 34 unrelated coupling). Both committed to `main` and deployed (v1.11.11, v1.11.12). Plugin version bumped twice during finishing. - RESEARCH Open Q1 (worker-perch asymmetry) carried forward as separate investigation — orthogonal to Phase 25.4 GOAL. - Operator approved close without third reprobe; the gate-predicate fix has positive proof (sentinel observed at nested path), the early-return fix has reasoned justification + downstream tests intact.