| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
(v4.6.161)
1. The last two `navigator.onLine`-as-fact payloads
- module-aiassist.js: the debug/health payload carried
`online: navigator.onLine`, consumed by downloads/dashboard.html, which rendered
"π΄ Offline" β on an Android WebView (reports false while fully connected) that
told working users they were offline. The key is gone; the payload carries
`network:'unknown'`, `networkKnown:false` and `networkHint`.
- module-world-model.js: the snapshot had the same field, and it is also handed to the
LLM through `_modelDigest()`/`ask()`, so the model was told the user was offline.
Now `network:'unknown'` + `networkKnown:false`.
- `_loadFromVFS()` scrubs `system.online` from a model persisted by an older build β
otherwise the stale boolean would survive inside the digest for the life of the VFS.
- downloads/dashboard.html: `networkText()` renders reachable/unreachable/http-* only
when doctor produced them ("(doctor)"), `unprobed` as "β not probed", and everything
else as "β unknown (run doctor for a live probe)". A legacy payload's `online:false`
lands in "unknown", never "Offline". The row label is "Network", not "Online".
- Neither payload invents a probe: SystemModule stores no last-doctor verdict, so
'unknown' is the honest value until one exists.
2. `vfsSize` in both payloads was a different unit again
- aiassist: `JSON.stringify(fs.root).length / 1024`; world-model:
`JSON.stringify(fs.root).length`. Same lie v4.6.160 fixed in `vfs-size` β UTF-16 code
units plus serialisation structure. Both now sum UTF-8 bytes of file content
(`_utf8Len`/`_measureVfs`, Blob β TextEncoder β code-point fallback, depth/node
capped) and carry `sizeBytes` beside the human string so the rows reconcile.
- world-model's `_gatherFileInfo().size` used a THIRD unit on an adjacent line
(`JSON.stringify(fs.root).length/1024`); now the same measurement, plus `sizeBytes`.
- `showModel()` renders exact bytes and tolerates a pre-fix persisted string.
3. Predictor made a guaranteed-401 request on every boot (UI-plan finding F6)
- `_wdConfig()` returns `user:''`/`pass:''` when the saved WebDAV config has a URL but
no credentials, and `_pullFromServer`/`_pushToServer` fetched the auth-gated
`/spaces/<user>/_data/predictor.json` anyway β 107 Γ 401 in one day, logged by the
browser as console errors. Both now refuse without credentials
(`cache-only (no credentials)`).
Verified: 42/42 node tests, 2 new files (38 payload/honesty assertions, 8 predictor-store
assertions); the predictor suite is 4/8 against the pre-fix source, and the payload suite
fails 14 assertions against HEAD (`online:false` present, vfsSize === JSON.stringify
length). Live Chromium on v4.6.161, on the served dashboard: a legacy payload with
`online:false` renders "Network β unknown (run doctor for a live probe)", the current
payload renders "85.8 KB (87880 B)" and the same unknown, and a doctor-produced
`unreachable` renders "β οΈ unreachable (doctor)". Served bytes identical for
core-bundle.js, main.html and downloads/dashboard.html. `maintain.py --fix` all PASS.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
phantom-delete fixes (v4.6.160)
Applies the three 2026-11-09 patches after review, and fixes two more
navigator.onLine lies found while verifying them. Each patch was checked
against the live code first; one of its claims did not survive.
1. virtual-fs β phantom deletes and checkpoint bloat
- deleteCheckpoint('nope') returned true and printed "Deleted checkpoint"
for a name that did not exist: _resolvePath() validates intermediate
segments only, so the leaf was never checked. New _requireLeaf() refuses
a resolved path whose leaf is absent, and deleteCheckpoint uses it.
- CHECKPOINT SIZE: checkpoint() cloned the whole root minus
_snapshots/_sessions/.trash, so _mods (module originals) and
.semantic_index were duplicated into EVERY checkpoint β measured
2,101,686 bytes for a 2 MiB _mods. Both are now excluded on write and
preserved across rollback (without the keep-list change a rollback would
delete the live module originals). Measured after: 185 bytes.
- Claim corrected: the patch also claimed deleteFile() reported success on
a phantom path. It did not β live code already required
`node && node.type === 'file'` and returned false. The guard is kept as
defence-in-depth, and the test asserts the real behaviour rather than the
claimed one.
2. module-fs β vfs-size reported ~2x the real size
- It measured JSON.stringify(root).length: UTF-16 code units PLUS every
structural character of the serialisation. A 1046-byte tree printed as
1961; anything under ~5 KB printed as "0.00 MB".
- Every number is now stored bytes (UTF-8 of file content): one measure()
walk for the total, the per-file table, the top-directory table and the
system dirs, so the rows add up to the total. The patch's arithmetic was
still off in seven places, all fixed here: a UTF-16 fallback that
reintroduced the bug when Blob was missing, a top-5 table that could
never sum to the total (now top 10 + an aggregated remainder row), the
top-dirs table dropping files stored at root (now a "/(root files)" row),
a 2-decimal MB total that printed 0.00 for small VFS, three duplicate
hand-rolled walks, readDir() used for dir sizes, and "1 files".
3. module-system β status/doctor no longer trust navigator.onLine
- status printed "Network: Offline" from navigator.onLine, which an
Android WebView reports false while fully connected.
- doctor now awaits a real HEAD probe (AbortController, 4 s): only
'reachable' is healthy, a throw is HIGH "Network: unreachable", a non-ok
answer is MEDIUM and names the status (405 to HEAD is not "offline").
- THE PATCH'S THRESHOLDS COULD NEVER FIRE: it compared localStorage bytes
against navigator.storage.estimate().quota, which is the whole-origin
quota (GBs, mostly Cache Storage), while localStorage has its own ~5 MB
cap β a 4.5 MB localStorage reported "All systems healthy" against a
2 GB quota. Now: localStorage measured as (key+value)*2 UTF-16 bytes with
HIGH > 4 MiB / MEDIUM > 3 MiB, VFS measured in UTF-8 bytes with
HIGH > 3 MiB and > 80% of the measured localStorage bytes. The origin
quota appears only in one informational line and decides no severity.
- Also removed navigator.onLine from version, bug-report and bug (a grep
for it is now 0 in the module), and added a "Measured:" line so a clean
report still shows what was measured.
4. terminal.js / module-servermon.js β the same lie, where the user looks
- The status bar printed "β Offline" in red from navigator.onLine; on a
WebView that is false while online. It now shows "πΆ Network?" with a
title pointing at `doctor` unless the browser claims online.
- `net` test line now reads "Browser reports: offline (unreliable)" as a
warning instead of a red failure β the real verdict comes from the
probes that follow.
Verified: 40/40 node tests (3 new files: 36 + 53 + 34 assertions, each proven
red against the pre-patch code; the doctor test runs against the PATCH's
threshold logic and shows the 4.5 MB case staying silent). An independent
orchestrator probe re-derived all of it: phantom delete false, 2 MB _mods
absent from the checkpoint and preserved through rollback, vfs-size exact
bytes for ASCII/Γ©/β nested trees, 4.5 MB localStorage HIGH despite a 2 GB
quota, failing probe unreachable, status never claiming Online/Offline.
Served bytes identical for core-bundle.js / main.html / module-system.js /
virtual-fs.js. Live Chromium on v4.6.160: vfs-size "85.8 KB (87880 bytes)
across 40 files", doctor "Measured: VFS 0.08 MB UTF-8 Β· localStorage 0.07 MB
UTF-16 in 14 keys", a stubbed failing probe reports HIGH unreachable, the
status bar reads "πΆ Network?" when the browser lies, deleteCheckpoint on a
missing name returns false, and the console has 0 warnings and 0 JS errors.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
(v4.6.159)
1. Tool arguments are coerced once, at the dispatch boundary (module-aiassist.js)
- _coerceToolArgs(name, a) at :1584 reads the tool's own JSON schema and, for every
declared string prop, converts number/boolean with String() and object/array with
JSON.stringify (never "[object Object]"); declared integer/number props accept
unambiguous numeric strings. null/undefined stay absent so the handlers' defaults
and _pathArg still own "no value given".
- Called at both dispatch boundaries: runTool (:2650) and runModuleTool (:3565 β
reachable from view_module and from ship_module -> test_module/health_check, which
have their own parse+switch).
- A JSON body of "null" or "[1,2]" parses to a non-object; both sites normalise to {}
before any handler reads a.field.
- The 26 handlers keep their `a.x || ''` reads: `|| ''` catches only FALSY, which is
how a truthy number or object reached .trim()/.split()/.replace() and filed one
auto-report per tool (user_get "key.split is not a function", create_skill,
patch_skill, delegate_task "task.substring is not a function", create_mindmap,
web_fetch, list_skills, ship_module).
2. gunzip is binary-safe (cryptogzip.js β plan 2026-11-03 T3, the last open task)
- _cmdGunzip used _vfsRead + _gunzipString (base64-only), so a real .gz fetched with
wget reported "not valid base64 input (illegal character \u001f)" and could not be
decompressed at all. It now reads bytes (_vfsReadBytes handles both the module's
base64 output and raw Latin-1 bytes), checks the 1f 8b magic, and reports
"not a gzip file" for non-gzip input without writing or deleting anything.
The module's own gzip output still round-trips.
3. Boot log is clean (terminal.js)
- Seven _safeRegister() calls for Editor/Shell/Alias/Finder/FileManager/Diff/BusyBox
were guaranteed no-ops: those classes are declared in critical-ext-bundle.js, a
<script> after module-loader.js, so they do not exist yet when new WebX11Terminal()
runs; adding them to the registry map would still resolve to undefined. Removed with
a comment β the ext bundle registers each one on load.
- SimpleGUIModule IS a core class but was missing from _safeRegister's registry map,
so the fallback could never restore it. Added.
4. New guard: admin/scripts/verify-plan-age.py + profile wrapper plan-age-guard.sh
(cron newos4-plan-age-guard, no_agent, daily 07:10). Silent when healthy; reports
plans marked <!-- autonomy: review --> with no execution stamp older than 72 h β
the process gap that let four fix plans sit for six days before v4.6.158. Read-only
by construction (148 real plans byte+mtime identical before/after a run), rc=2 when
the directory cannot be scanned, --strict opt-in so a report is never recorded as a
crashed guard.
Verified: 37/37 node tests, 3 new test files and 631 new assertions (613 arg-coercion,
18 gunzip); test-cryptogzip-guard's C8 updated because it pinned the old misleading
"not valid base64" message. Independent orchestrator checks: all 66 declared string
props across 58 tool defs coerced under 4 hostile shapes; positive cases untouched
(string content not double-encoded, numbers still numbers, null still absent, unknown
tool identity preserved). Served bytes byte-identical for core-bundle.js, main.html and
cryptogzip.js. Live Chromium on v4.6.159: write_file with object content wrote 27 chars
of readable JSON (not "[object Object]"), read_file with offset:"0"/limit:"200" coerced
to numbers, user_set with key 12345 -> "12345" and object value -> JSON, and the boot
log now carries 0 SafeRegister warnings (52 messages, 0 JS errors) with SimpleGUIModule
registered.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Each item is symptom -> cause -> fix, re-verified against the served bytes.
1. debug-ai panel threw "Cannot read properties of undefined (reading 'substring')"
(RPT-20261108-231622). _updateDebugPanel and two sibling previews called
JSON.stringify(entry.args).substring(); JSON.stringify(undefined) returns
undefined, and _toolCallLog entries carry args: undefined for any tool call
with no arguments. All three previews now go through _formatToolArgs(), which
renders "null" and never throws.
2. test_module threw "Cannot convert undefined or null to object"
(RPT-20261108-232127). Object.keys(instance.commands) assumed every module
initialises a command map. The throw landed after loadModuleFromCode had
already succeeded, so the rolledBack:true branches were skipped and an
unverified module stayed loaded. Both sites now require instance.commands.
3. list_files threw "originalPath.replace is not a function" (RPT-20261109-172742,
3 occurrences in errors.json). A model can send a non-string path despite the
schema. The path family (list_files/read_file/write_file/make_directory/
delete_file) now coerces through _pathArg() and answers the tool's own honest
error instead of a TypeError.
4. wpm search failed for EVERY keyword (RPT-20261109-123907): packages.json entry
`kanban` used `description` while the other 31 use `desc`, and search() called
pkg.desc.toLowerCase() mid-loop. Catalog reads now go through _pkg() /
_packageMap() (desc || description, values coerced to strings), and the kanban
entry carries name/version/desc/commands. Re-running the original loop for
pdf/git/zzz is clean.
5. wx/webxai/webx turned a chat-lock timeout into an unhandled rejection
(RPT-20261107-150726): forwardToAI() called chat(args) in a synchronous
function without returning or catching the promise, so the 10 s busy throw
never reached dispatchCommand's catch. It now returns the promise with a
.catch that prints "AI error: ..." through terminal.print β console.error is
hooked by the debug supervisor and would re-file the same report.
6. CryptoGzip._base64ToBytes threw a raw DOMException on non-base64 input
(RPT-20261102-203616). It now validates first and throws
Error("not valid base64 input (<reason>)"). The duplicate shadowing
definitions at the bottom of the class were deleted β the later copy is the
one that carried the unguarded atob().
Also regenerated modules/core/core-bundle.js (generated, never hand-edited) and
bumped the build v4.6.157 -> v4.6.158 (main.html x3 + module-loader.js).
Verified: 35/35 node tests, including 6 new regression tests each proven red
against the pre-fix source; an independent orchestrator harness 36/36; and
newos4-maintain.py --fix all PASS after regenerating the stale changelog page.
Served bytes are byte-identical for the bundle, main.html and both ext modules,
and a live headless-Chromium run against v4.6.158 shows WpmModule.search('pdf')
returning 1 result, the debug panel rendering a no-arg entry as "null",
_base64ToBytes('grΓΌΓe') raising a plain Error that names base64, and 0 JS errors
on load.
m.pull (RPT-20261102-211438) is closed as a usage error: no module in the tree
defines pull(); the real interfaces are the webdav-pull and model-pull commands.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
are explained
Symptom: operator reports newOS4 "started to crash since the last ui update"; a
user (Max) reports his browser crashed. Order: roll back, do not re-enable until
it is verified in a dev environment.
Root cause: not yet established, and the log work rules out the obvious suspect.
- 0 requests for modules/ext/ui-*.js in the entire access log: no interface mode
has ever mounted in production, so a mode's own chrome cannot be the cause.
- all six modes driven on one page in headless Chromium at 390x844: no crash, no
page errors, one #terminal-container, heap flat at 10 MB over a 25 s idle.
- the six errors.json entries that read like a crash report (SPA is not defined,
DOM harness FAILED: N, "...reading 'makeInstance'") exist in no shipped file -
grep on the served 1.29 MB bundle returns 0 - so they came from a harness run
inside that one browser session.
Two real defects were found instead, both planned not bundled here:
- /ollama/api/chat inherits nginx's 1 MB default client_max_body_size and
refused a 1,063,595-byte agent context with 413 (only 413 in 7 days).
- module-aiassist.js:2576 const originalPath = a.path || '' then .replace on a
model-supplied non-string: TypeError, 3 events on tool:list_files, 32 sites
with the same shape.
Fix: the UIShellModule registration in terminal.js is commented out with a dated
note and the revert recipe; bundles rebuilt from source; LATEST_VER v4.6.156 ->
v4.6.157; main.html's three ?v= refs bumped. The six modes are lazy ext files
whose first line is `if (!terminal.uiShell) return;`, so an accidental load is a
no-op and no first-run picker appears.
Verified: served core-bundle.js?v=v4.6.157 is cmp-identical to disk (1,298,695 B)
and the registration exists only as a comment (core-bundle.js:4082). Playwright
390x844: shellPresent false, uiShellModuleRegistered false, uiRoot false,
pickerNode false, moduleCount 47, help renders 21,802 chars, pageerrors [].
Workaround: none needed - users get classic only. Re-enable = restore the line,
rebuild, bump LATEST_VER + main.html, re-run admin/scripts/probe-ui-modes.py.
Plan: .hermes/plans/2026-11-09_224409-newos4-ui-rollback-dev-verify.md
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
two cards; commander was black
Three defects reported against v4.6.155, each reproduced before touching code:
1. 'the modules tab keeps opening' (workspace). api.capture() runs dispatchCommand() for real, and
the Modules card called bare 'modules' β whose only effect is terminal.openModuleManager().
Every 15 s refresh re-opened the panel (measured: #module-manager display:flex at t+9s and
t+18s). The card now captures 'listmods' (same inventory as text), and api.capture() refuses
UI-only commands outright so a future card cannot reintroduce it.
2. 'the picker got reduced to 4 items'. showPicker() used this.list() once four modes were
registered β but modes are lazy, so that is 'modes you have already visited'. It now always
renders the six in level order, taking registered specs for label/tagline.
3. 'commander is mostly black' on a phone. The base sheet's #ui-root rule used
body[data-ui-mode]:not(...) #ui-root (1,2,1) and therefore beat every mode's own
body[data-ui-mode='x'] #ui-root (1,1,1): commander/studio/cockpit asked for display:grid and got
flex column. Commander collapsed to a 253 px centre column in an 844 px page with both rails
hidden. Base specificity is now (1,0,0) with an explicit !important hide for classic/boot, and
the commander phone query pins grid-template-rows: minmax(0,1fr) so the single row fills.
Measured after: picker 6 cards after visiting four modes; #module-manager stays display:none
across a refresh cycle; commander display:grid, phone centre 390x810 with a 673 px terminal,
desktop 232/788/260 columns. probe-ui-modes.py --cycle -> ALL PASS, 13 checks; shell test 25/25.
v4.6.156
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
dedup, sweep race, a11y
Four models (kimi-k2.7-code, deepseek-v4-pro, minimax-m3, glm-5.3) reviewed the shipped change
set. Review log with every accepted and rejected finding:
.hermes/plans/2026-11-07_163420-newos4-six-ui-modes.review-log.md
- identity guard now covers every snapshot node, not just #terminal-container: terminal.js also
binds #output, #command-input, #input-area and #status-bar, so a cloned one leaves the terminal
writing into a detached node while the page looks perfect
- _restoreStock anchors on the next stock node already home, then on the recorded trailing
sibling; a raw child index was measured wrong (terminal,input,status-bar,quick-actions)
- _loading map: a double-clicked card or a repeated 'ui easy' loads the mode once
- _sweepPending returns while that id is in flight instead of declaring failure at 6 s and
overwriting the stored preference
- _syncChrome() on every switch (quick-actions was left in the tab order behind the mode root)
- picker: role=dialog/aria-modal/labelledby/focus, shown after 600 ms instead of waiting for a
registration that never happens with lazy modes, and suppressed when ?ui= already applied
- ?ui= must be one of the six; _loadMode only accepts packages['ui-<id>']; hasOwnProperty on modes
- VFS names are shell-quoted before reaching api.exec (cat/edit)
- refresh generation no longer bumps per tick (slow captures left later panes at '...')
- easy: padding shorthand reset the 34 px footer reserve (footer overlapped the composer)
- cockpit panes get tabIndex and focus; F1/F12 left to the browser; nothing swallowed in inputs
- api.el lost its unused innerHTML sink; the test lost its _parseArgs stub (it hid the regression)
- v4.6.155
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
handling
Reviewed the shipped change set with kimi-k2.7-code, deepseek-v4-pro, glm-5.3 and minimax-m3.
Findings below are the ones that survived verification against file:line; the rest are in
.hermes/plans/2026-11-07_163420-newos4-six-ui-modes.review-log.md
Accepted and fixed:
- apply() had no generation token: a lazy mode load that finished after the user switched to
another mode force-applied the abandoned one (`apply(want,{force:true})` in the .then).
Every apply() now bumps `_loadSeq` and the callback acts only while its token is newest.
Same token also removes a double mount: registerMode() already applies the mode as it
registers, so the callback was mounting every lazy mode twice (teardown+cleanup twice).
- _restoreStock() used the recorded `nextSibling` as insert anchor. Every real mode moves
SEVERAL stock nodes (terminal, input, status bar), so the recorded sibling is inside the
mode container and every insert fell through to append. Now anchors on the next stock node
that is already back home. A raw child index also fails (elements removed shift the
indices) β measured order terminal,input,status-bar,quick-actions.
- _abortToClassic() left the broken mode persisted and `_pending` set, so the next reload
retried the same failure with no way out. It now clears both and stores 'classic'.
- ?ui= reached classList.add('wx11-ui-' + id) unvalidated: a space or colon throws a
DOMException and boot dies. Validated against /^[a-z][a-z0-9-]{0,15}$/ and no longer
persisted (the comment always claimed it was a request, not a preference).
- Modes hide .prompt-label, so #command-input had no accessible name at all. The shell sets
aria-label once when a mode reparents it.
- cockpit: preventDefault on all F1-F8 killed help, the address bar and devtools. F1/F12 are
left to the browser and nothing is swallowed when focus is in a text field.
- commander: Ctrl/Cmd+K opened the palette even while typing in the composer. Ignored from
inside editable elements now.
- picker: role=dialog + aria-modal + aria-labelledby + focus on the first card.
- api.capture(): strip the quoting characters from argv (capturePipeOutput takes argv).
- stale `wx11-ui-*` body class removed on switch instead of accumulating.
Rejected with evidence:
- "loadModuleFromURL may not return a boolean" β modules/core/module-manager.js:56 returns
true on success and :60 false in the catch. Verified by reading the function.
- "_abortToClassic does not clear _pending" β it did not; that one is accepted above. The
sibling claim that it "leaves the broken mode persisted" was verified by reading the
function: it never touched localStorage, so also accepted.
- "mm.LATEST_VER / webx11_cache_ver are never set, so the mode URL falls back to ?v=0" β
modules/core/module-loader.js:12 sets window.webx11.moduleManager.LATEST_VER and :14-15
writes webx11_cache_ver. Verified live: the mode fetch carries ?v=v4.6.152.
- "moduleManager.modules may be a Map" β it is a plain object (terminal.moduleManager.modules
.CryptoGzipModule is the documented access path), so ui-commander's Object.keys works.
Verified: node modules/test/test-ui-shell.js -> 25/25 (the test now moves TWO stock nodes; a
raw-index restore fails it with "terminal-container,input-area,status-bar,quick-actions");
probe-ui-modes.py --cycle against production -> ALL PASS, 13 checks.
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
welcome modal
- #ui-root is a flex column (it was display:block, so every mode's flex:1 was inert and the
terminal collapsed to 0px while the composer sat at the bottom of a 4600px flow)
- overflow:clip on #ui-root and body: overflow:hidden is still a scroll container, and the
terminal's scrollIntoView() on new output scrolled the mode's own chrome off-screen
- terminal min-height 100px: in the tabbed/panel modes its grid parent is not a flex parent
- api.capture() split arguments itself; term._parseArgs does not exist, so every panel showed
'unavailable: term._parseArgs is not a function'
- the classic Let's Get Started modal is suppressed in non-classic modes: it covered easy's cards
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Symptom: newOS4 had exactly one layout. A first-time user had no way in that did
not require learning commands, and an operator had no surface showing live runtime
state; the only way to see heap/VFS/tool-call telemetry was to type cat /sys/...
Root cause: no UI layer existed. main.html hard-codes one flex column, and
terminal.js binds the stock nodes by id at construction, so an alternative layout
had to either duplicate the output stream or reparent the real nodes.
Fix:
- modules/core/module-uishell.js (new): mode registry, #ui-root, stock-node
snapshot/restore, identity check that reverts to classic when a mode removes or
clones #terminal-container, shell-docked #ui-foot (the AI stop button exists in
every mode), ui command, gated first-run picker, ?ui= honoured only if the mode
actually registers, lazy mode loading via the same loader ext modules use.
- modules/ext/ui-{easy,chat,studio,workspace,commander,cockpit}.js (new): six
reparenting skins with scoped CSS. Workspace/commander/cockpit read real /sys/*
mounts and captured command output; no value on screen is invented.
- modules/test/test-ui-shell.js (new): 25 assertions - teardown, clone rejection,
failed-mount fallback, footer docking, destructive-mode rejection.
- terminal.js: register the shell; make one failing ext module stop aborting the
rest of the serial load loop (it did, silently, which is why six ui-* files at
the end of extKeys never loaded).
- module-loader.js: LATEST_VER v4.6.152. The six ui-* keys are deliberately NOT
in extKeys - see the rate-limit note below.
- admin/scripts/probe-ui-modes.py (new): 6 modes x 2 viewports plus a --cycle pass
that switches all seven states on ONE page and asserts the output node keeps its
identity.
Also fixed, found while testing (pre-existing, production-affecting):
- nginx limit_req zone=webx11 burst 50 -> 200. One newOS4 boot fans out ~40 assets;
the bucket emptied mid-boot and every later module got 503. Measured in
webx11.error.log: "limiting requests, excess: 50.696 by zone webx11" against
modules/ext/ui-chat.js. A reload within ~50 s silently broke modules. Sustained
rate unchanged at 60r/m.
Verified: node modules/test/test-ui-shell.js -> 25/25; removing both _restoreStock()
calls measures 20/25 with five FAILs; probe-ui-modes.py --cycle against production ->
ALL PASS, 13 checks; served bytes: v4.6.152 x3 in main.html, class UIShellModule x1
in core-bundle.js, six ui-* packages in packages.json.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
S1 (reported by agent-os4-maint-01, wrongly refuted by the chat auto-reply):
wget existed but was unusable and dishonest.
- Cause 1: _cmdWget used the page's own fetch(url), so every cross-origin URL
died on CORS ("Failed to fetch"). Measured in the served v4.6.150 page:
fetch(cgit README) threw, /api/fetch returned 200/1954 bytes.
- Cause 2: _vfsWriteBytes ignored writeFile()'s boolean, so the command printed
a green tick when the target directory did not exist. Measured:
fs.writeFile('nosuchdir/x.txt') === false.
- Cause 3: bytes were round-tripped through String.fromCharCode, mangling UTF-8.
- Fix: fetch through /api/fetch (raw source), mkdir -p, writeFile + read-back
verification, error instead of a tick on failure. Register curl with
-s/-o parsing delegating to the same writer so the two cannot drift.
S2 (same reporter): web_fetch could not page and said it could.
- Cause: the handler honoured max_chars up to 50 000, then _clipToolResult kept
toolResultChars-200 = 11 800, so 38 200 chars vanished silently; offset was
not in the schema and was never read, so next_offset pointed at a page the
caller could not request again.
- Fix: offset/next_offset/total_chars/source_bytes in the contract; the page is
built, measured against the clip budget and shrunk until the envelope fits,
so next_offset - returned_chars is always 0.
- Also fixed: _clipToolResult appended "0 of N chars withheld" when the content
already fitted, claiming a loss that did not happen.
C1: fetch <url> --save=FILE β the only terminal path from a remote URL to the
VFS, so a 90 KB source page no longer has to be read through 3 000-char
scrollback.
Proxy: /api/extract now returns source_bytes (raw page size) and
content_chars_utf16. Python len() counts code points while JS String.length
counts UTF-16 units; 7 astral characters in the changelog left a 7-char gap at
the end of a 90 KB page (measured). _html_ok declares charset=utf-8.
Verified: node --check on all four modules; py_compile; bundles rebuilt;
served bytes identical to disk for all five assets at ?v=v4.6.151; live in the
browser: curl -s -o review/main.md <cross-origin> saved 1954 chars with the
correct head, a 404 wrote no file, fetch --save wrote 91 518 chars,
web_fetch p1 returned 11 754 with next_offset 11 754 (gap 0), page 2 started at
a different offset with next 23 496.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Operator commit-approve, 2026-11-06. These changes have been served since
2026-11-02/03 while git HEAD still stamped v4.6.146, so the mirror published a
tree four builds behind the running instance and the release tag named the
wrong build.
module-orchestrator.js β orchestrate-parallel reported success it could not
prove. Results are now recorded by sub-task id through a resultSink handed to
delegateTask (harvest matched by currentTask/lastTask strings, which a later
delegation on the same agent overwrote, losing finished work as "agent not
found"); the tally counts summaries with status ok instead of
Promise.allSettled resolutions (a spawn refusal fulfils with undefined and used
to print green); and a busy agent is no longer reused at pool cap, where two
siblings erased each other's lease. Guard:
modules/test/test-orchestrator-parallel-harvest.js (16 cases, reproduces the
one-agent three-sub-task cap case that printed 3/3 succeeded).
module-billing.js β the balance said "not signed in β newOS4 sends no session
token", which was already false: the browser sends the HttpOnly cookie and
parse_bearer_token() falls back to it. `cost` is on by default (the flag was
opt-in and nobody knew the command existed; `cost off` opts out) and the
command now prints today's platform spend from GET /api/billing/usage?days=1
alongside the session estimate.
modules/ext/webdav.js β the inline WebView detector missed "Version/4.0"
Android WebViews, including the operator's own phone, so the login offered a
Telegram widget that cannot complete there. It now delegates to
modules/wx11-tg-webview.js when the page loaded it and keeps a matching
fallback (wv token, Version/4.0-Android, Telegram, FBAN/Instagram/Line).
modules/ext/media.js β text analysis model renamed deepseek-v4-flash ->
deepseek-v4.1-flash in both the field and the printed label.
main.html + module-loader.js β v4.6.146 -> v4.6.150 in LATEST_VER and the three
asset queries; core-bundle.js rebuilt (rebuild-bundles.py reports it already
in sync, 1,250,569 bytes).
Verified: node --check on all six scripts and the new test; the 16-case
orchestrator guard passes against the deployed bundle (ORCH_SRC=...); bundle
rebuild is a no-op; core-bundle.js?v=v4.6.150 answers 200.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Two dashboards existed and had drifted apart: root /dashboard.html (118 KB single file with
Subscription & Usage + API Keys) and /spaces/dashboard.html (34 KB shell + assets/js/dashboard-*.js).
The nginx log settles which one is real: 77.211.5.129 requests /spaces/dashboard.html and never
the root page. Tonight's fixes landed on the root page twice before that was noticed.
Neither page was a superset, so deleting outright would have lost features:
root only -> Subscription & Usage card, API Keys card (create/revoke/reveal)
spaces only -> 2FA/TOTP, change password, delete space, email change, files, logs
assets/js/dashboard-billing.js + assets/css/dashboard-billing.css (new)
loadSubscription / loadApiKeys / cancelSubscription / createApiKey / revokeApiKey /
copyNewApiKey / dismissNewKey + isProPlusPlan / capPlan, lifted verbatim from the root page,
wrapped in an IIFE that re-exports the onclick handlers on window; getAuthHeaders() reads the
modular app's localStorage.webx11_session_id; PRO_PLUS_PLANS carried over (its absence threw
"PRO_PLUS_PLANS is not defined" and silently hid both cards on the first attempt)
spaces/dashboard.html
the two cards inserted after the Activity Logs card, plus the css/js includes
dashboard-auth.js showDashboard() calls loadSubscription() once the view is up
retirement
root dashboard.html deleted (git rm); `location = /dashboard.html { return 301
/spaces/dashboard.html$is_args$args; }` so bookmarks keep working and ?next= survives
links repointed in ai-api.html (6), ai-course.html (3), models.html (1), newOS4 hint string
(v4.6.146 + rebuilt bundles so the cached bundle cannot serve the old URL)
Measured (playwright, 390x844 WebView UA, agando's real session):
subscriptionCard display block, height 315px, plan "Free", price "EUR0/mo", badge "Free",
usage "$0.00 / inf", API-keys card hidden for a free plan, 0 console warnings
/dashboard.html?next=/skills.html -> 301 -> /spaces/dashboard.html?next=/skills.html, 200,
login card + Telegram button present
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Two hosts cannot complete the widget flow and both are in use: an Android WebView /
in-app browser (the operator's phone UA ends "; wv)") and a desktop browser that blocks
the popup or hands the link to the Telegram app. The widget needs a popup that can
postMessage back into the opener; in those hosts the link lands in a browser tab or in
Chrome with no opener, onTelegramAuth never fires, and the login card looks dead β
"it opens my default browser" (reported twice: phone, then Max's laptop).
modules/wx11-tg-webview.js (new)
WX11TG.isWebView() -> "; wv)", Telegram in-app, Meta in-app UAs
WX11TG.dropWidgets()-> removes already-rendered oauth.telegram.org iframes/scripts
dashboard.html
code box is open by default and drops the nested <details>
WebView: widget removed + "cannot work inside an app window" hint
desktop: if the widget was clicked and onTelegramAuth has not fired in 12 s, open the
code flow with a hint instead of leaving a dead card
spaces/newos4 (webdav-login panel)
in a WebView the widget is never injected, the mount is hidden and the device-code
flow starts by itself; desktop keeps the widget unchanged
version v4.6.145
Measured with playwright (390x844 WebView UA and 1440x900 Chrome):
WebView -> 0 Telegram iframes, hint shown, code box open, real WX11- code minted
desktop -> widget still rendered, code box open, real WX11- code minted
newOS4 WebView -> no widget injected, code area auto-visible, code WX11-..., no JS errors
newOS4 desktop -> widget still offered
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Third occurrence of the same class of bug, and the first fix is why: commit e95b7be
(2026-11-01, "multi-session .newsession v3") touched exactly ONE hunk of
auth-file-api.py β handle_login. handle_device_poll, handle_telegram_login and
handle_register kept overwriting the whole .newsession with a single-session v2
payload, so a phone sign-in logged the desktop out and the loser held a session id
no store knew: permanent `401 Invalid session or API key` (newOS4 asked for a login
it could never complete). Measured: agando's store held 1 entry (sha8 45cb7819) while
the phone sent 9ac607f1 for 3 h; 26 device-polls in 2 s from one page minted 26
cookies and only the last write survived, which also drained the shared nginx
`webx11` zone (11x 503 on /admin/api/*).
api/webx11_auth.py
- add_session() is now THE writer: v3 `sessions[]` (cap 5) under a per-user
`flock` (LOCK_DIR outside the spaces), plus the legacy `.session` mirror
- module-level import so no handler has to reach for it lazily
api/auth-file-api.py
- device-poll / telegram-login / register / recover-by-backup-code all call
add_session(); a bare write_newsession_file() is now only the admin-gated
handle_create_session
- handle_device_poll runs its approved branch under a per-code `flock` and mints
exactly one session per approved code (idempotent under a poll storm), with a
cheap pre-lock reject so unvalidated codes cannot mint lock files
- handle_me_cookie expires an unknown session cookie (Set-Cookie Max-Age=0,
code SESSION_DEAD): the cookie is HttpOnly, so no page could ever clear a dead
credential and every load re-sent it
api/newcrypto.py
- write_newsession_file() is atomic (tmp + fsync + os.replace); the temp name
carries pid+tid because a fixed `<file>.tmp` lets two concurrent writers of one
store steal each other's temp file (reproduced: FileNotFoundError)
admin-login.html single-flight poll (polling/finished guards) β 26 calls -> 1
dashboard.html device-code fallback when the Telegram popup has no opener
spaces/newos4/* 401 -> session-repair notice + endpoint normalisation, v4.6.144
nginx /api/auth/ + /api/security/ on zone wx11auth (120r/m, burst 60)
admin/scripts/verify-session-writers.py NEW static guard, wired into smoke-auth.sh
Proof (live, after restart): 5 concurrent polls on one approved code -> 5x approved,
1 distinct session id, store v3 with 2 entries, pre-fix session still valid; dead
cookie -> 401 + Set-Cookie expiry, valid cookie -> 200; 200-call auth burst -> 503s
only on /api/auth/, /admin/ stayed 302; smoke-auth.sh PASS. Guard RED (3 offenders)
-> GREEN. Backups: /home/darkbot/backups/authfix-20261102_214424
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
invents "completed"
Two defects found by re-deriving a report that claimed both were still open.
read_file paging (module-aiassist.js)
- A page larger than the result clip was clipped AFTER read_file had computed
next_offset. On the last page next_offset is null, so the model was told
truncated:false while its tail had been removed: measured 11 890 of 19 999
chars delivered, returned:19999, no pointer to the rest - 8 109 chars lost
with no way to ask for them.
* the clip now sets truncated=true and next_offset=offset+keep when it removes
the tail of a page that had no continuation (and never splits a surrogate pair)
* read_file caps a page at the result budget (min(pageCap, toolResultChars-200)),
so a page is always delivered whole and next_offset is exactly contiguous
Live v4.6.143: 19 999-char file -> page 11 800 (next_offset 11 800), page 2
8 199 (next_offset null): 11 800 + 8 199 = 19 999, nothing withheld. A model
asking for limit:32000 is clamped to 11 800 and told truncated:true.
The ~500-char symptom was the 500-char floor: a caller passing limit:200 gets
500, which is why targeted sed -n reads looked better than the tool.
Delegation reporting (module-orchestrator.js, module-aiassist.js)
- chat() returning {content} was already fixed in 8a9333b, but three other places
still substituted the word "completed" for an absent answer, so the symptom
survived: delegateTask's outbox wrote status:'completed' even when the worker
returned nothing; the delegate_task tool returned {success:true, result:"completed"};
orchestrate-parallel summarised an unmatched agent as {status:'ok',
result:'completed (agent recycled)'} and a missing outbox as "completed (no outbox)".
- Now: an empty/undefined answer is status 'failed' with error 'agent produced no
answer' / 'no result in outbox' / 'agent not found β result unverifiable', and
the tool returns success:false with result:null.
Tests: test-delegation-reporting.js 6 (RED 3/6 against HEAD: D3/D4 success:true,
D5 the literal "completed"), test-tool-context-bound 26 with E6-E8 (RED 24/2:
truncated:false, next_offset null). maintain 27/27.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
unknown
Three follow-ups from the merge review, each with the assertion that pins it.
module-memory-graph.js
- _normalizeEdges no longer mutates stored weights. It capped legacy 'failed' edges
from 0.8 to 0.2 in memory, and the next debounced save made that permanent - an
irreversible migration of the user's graph with no schema version. The cap moved to
_effectiveWeight(), which ranking (showStats "Strongest relationships"), the session
view and prune() now use; the stored record is left exactly as written.
- config.json is stamped version: 2 so a future format change can migrate deliberately.
ext/webdav.js
- the two fallback parsers emit size: null instead of 0. They cannot see the size
column, and 0 rendered as "0 B" - the same "every file looks empty" symptom as the
parseInt bug fixed in e639c2d, one regex miss away.
- _formatSize(null) renders 'β' (unknown) rather than '0 B'; 0 still renders '0 B'.
module-aiassist.js
- the lock timeout now says what to do: "AI is still busy with another turn (waited Ns).
Wait for it to finish, or run 'ai-stop' and retry - the running turn's lock is left
alone." It used to be a dead-end string.
Tests: test-memory-graph 39 (T6.4 rewritten to the read-time contract: stored weight
stays 0.8, effective 0.2, success unaffected), test-webdav-sizes 28 (A14d null -> 'β').
maintain 26/26, live v4.6.141: _effectiveWeight 0.2 with the edge object unchanged,
_formatSize(null) 'β', '0 B' still '0 B'.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Closes STEER-20261102-205416-facd and STEER-20261102-204031-f82a.
WebDAV (modules/ext/webdav.js):
- listings showed "21 B" for a 21 KB file: nginx emits humanised sizes
(autoindex_exact_size off, vhost :412) and _parseAutoindex did parseInt("21K").
_parseAutoindexSize() handles K/M/G, decimals, lowercase and "-"; PROPFIND's
exact getcontentlength path is untouched.
- webdav-pull into a missing local directory printed "Pulled: ... (N bytes)" and
wrote nothing: it ignored fs.writeFile's false return (virtual-fs.js:561).
It now creates the parent chain, checks the write, and only reports success
after a confirmed write. The byte count uses TextEncoder, not .length
(a Cyrillic file under-reported ~2x).
- _formatSize gained the GB branch the parser already accepted (3 GB printed
"3072.0 MB") and no longer prints "NaN MB" for a missing size.
MemoryGraph (modules/core/module-memory-graph.js):
- addEdge averages weights, so every repeat pulled a strong edge toward the
call site's constant (0.9 -> 0.7 -> 0.6). Now Math.max, with decay left to
applyDecay(); prune() untouched.
- task ids were `task-${Date.now()}`, so re-submitting a task never deduped:
now a hash of agent + normalised text.
- recordFileAccess no longer records the graph's own /.memory_graph/ store.
- failed tool calls were weighted 0.8 vs success 0.5, so showStats ranked
failures as strongest: now 0.2. Legacy 0.8 failures are capped on load.
- concept ids keep punctuation ("concept:(v2!),"): slugged, and the slug uses
\p{L}\p{N} so Cyrillic/CJK/accented text still yields nodes -- an a-z0-9 slug
silently produced none.
Review findings folded in from the regression pass: the non-Latin regression
(a-z0-9 slug), the GB formatter asymmetry, and the pull byte count were all
found by review, not by the patch authors, and each now has a test.
Tests: test-webdav-sizes.js 27 (was RED 15/24 before the fix), test-memory-graph.js
37 (RED 17/32), both auto-discovered by newos4-maintain -- 26/26 node tests.
Live on v4.6.140: _parseAutoindexSize('21K') 21504, _formatSize -> "21.0 KB",
_formatSize(2GiB) "2.0 GB", slug Cyrillic "ΠΈΡΠΏΡΠ°Π²ΠΈΡΡ-ΠΏΠ°ΡΡΠ΅Ρ-ΠΎΡΠΈΠ±ΠΎΠΊ", 0 console errors.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
a lock
Closes the second half of STEER-20261102-201045-ab08.
- chat() gated the main-turn lock on nothing, so orchestrate-parallel's fan-out
(each sub-task calling chat() on the same instance with its own agentTag) ran
strictly one at a time. Measured 2x300ms delegated runs: 617ms before, 306ms
now; the two 200ms user turns still take 400ms (main path stays serialised).
- the 10s lock timeout called this._chatLock = false and printed "force
released", which cleared a lock the waiter did not own - that released the
main turn's lock while it was still running. A waiter now throws and the owner
is the only writer.
- delegated runs no longer clear the user's cancel flag on entry (an agent start
used to reset this._cancelled, disarming a Stop the user had pressed).
- tool-log entries carry the run they belong to and _recentToolLog(tag, n) is
what feeds the run log line and _postRunReview, so concurrent runs no longer
consume each other's entries.
Known limit, unchanged: _currentTag is a single field, so under true concurrency
an entry can be attributed to the wrong run (telemetry only - the answers are
per-session). Namespacing it properly needs per-call state threading.
Test: modules/test/test-chat-concurrency.js, 7 assertions, RED before
(H1 617ms, H3 force-released, H5 no tags) and GREEN now; maintain 24/24.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Regression from 32608bc, caught by the .chat() caller sweep before it aged:
- `recent` was referenced in chat()'s finally block but never declared, so any
turn with toolTurns > 0 threw ReferenceError there and chat() REJECTED.
delegateTask's catch turned that into "agent failed"; the 'completed'
fallback was only ever reachable on tool-free turns.
- chat() still had no return at all, so a tool-free turn resolved undefined and
module-orchestrator.js:1250 fell through to `|| 'completed'`.
Both fixed: `const recent` is now built from _toolCallLog, and chat() returns
{ content: finalAnswer, toolTurns } before the archive block (a return at the
end of the method would let _archiveSession swallow the answer).
Verified with /tmp/probe-chat-hotfix.js, 6/6: tool-using turn resolves
{content:'WORKER ANSWER', toolTurns:2}, tool-free turn resolves the answer,
the orchestrator's own projection returns the answer instead of 'completed',
and _chatLock is released. Run against 32608bc the same probe says
"recent is not defined" - i.e. it reproduces the shipped defect.
Not included: the _chatLock serialisation (orchestrate-parallel still runs
sub-tasks one at a time and the 10s force-release can clear the main turn's
lock). That stays open in STEER-20261102-201045-ab08.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Closes STEER-20261102-201234-7200 (read_file capped at 6000 while the schema
promised 32000).
Budgets (module-aiassist.js + new modules/core/model-profiles.js):
- one resolver, precedence: call arg -> loop profile -> model profile ->
ai-config -> code default. Keys: read-page-chars, read-result-chars,
tool-context-budget, fetch-chars.
- model classes tiny/small/mid/large keyed by id then by size token; `mid`
reproduces the old numbers so nothing moves for a mid model.
- _clipToolResult now clips the `content` field, never the serialised JSON
envelope: clipping the envelope deleted truncated/next_offset, so a 16k page
lost its continuation pointer (the reported symptom).
- read_file's schema no longer promises a number the model class may not give.
Telemetry:
- _recordToolCall keeps sizes, duration, ok and a 120-char arg preview instead
of the whole payload; ai-stats prints calls / ok% / avg ms / KB / truncation
from localStorage webx11_tool_stats.
Server (api/proxy-api.py):
- /api/extract honours max_chars (default 12000, cap 100000) and returns
content_chars + truncated; the bs4 branch is no longer uncapped while the
regex fallback was silently cut at 10 000.
BEHAVIOUR CHANGE: num_ctx for a model with no explicit setting now follows the
class default (16384 -> 65536 for cloud/large, 8192 for tiny). Cloud models bill
on tokens used, so cost is unchanged; local models get a bigger KV cache.
Two build traps, both now documented in AGENTS.md:
- model-profiles.js must be the FIRST entry in CORE_FILES and must export the
MODEL_PROFILES global (a TDZ typeof throws; without the global every budget
silently falls back to the code default).
- a budget key needs the setter branch AND the constructor loader at
module-aiassist.js:49 - persisting without restoring looks fine until reload.
Verified: test-tool-context-bound 23/23 (RED before: E1-E4 + TypeError on
_toolLimit); live bundle read_page 32000 / result 12000 / budget 48000 /
fetch 12000 and _ctxTokens 65536 for gemma4:31b-cloud; /api/extract 28970 ->
500 chars with truncated true; ai-config write+restore probe 6/6.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Operator policy 2026-11-01: every subscription may call any model, and the free tier resets daily.
Model gate removed (api/ai-gateway.py)
`plan == 'free' and model_weight > 1 -> 403 "Advanced models require Pro"` ran before any money
logic, so a free-plan account holding $1086 of purchased credit was refused and told to upgrade.
A model tier is not something we sell: the allowance and the USD balance bound spend. A caller who
cannot pay now gets the honest 429 (with period + needed amount). model_weight survives only for
track_usage() analytics.
Free allowance: $0.50/month -> $0.10/day
~$3/month retail = $2.14/month of provider cost at 1.4x markup, ~6x the previous giveaway and
still bounded per member (the 50-slot invite cap bounds the total; at full occupancy ~$107/mo or
$2.14/user at 100% utilisation).
The window is policy, not code:
- admin/data/pricing-policy.json: free_tier_usd=0.1, free_tier_period="day"
- wx11_billing: FREE_TIER_PERIOD, allowance_window(plan) (day for free, month for paid),
period_start(window), spend_usd(session, window) β one definition used by BOTH the gateway and
the billing API (the old month_spend_usd duplicate is what let them drift before)
- gateway + /api/billing/subscription now report the same window; plan_allowance gains `period`
- dashboard.html, newOS4 module-billing, chatstudio account panel say "today" when period=day
- published copy: index.html, pricing.html (card + compare table + how-usage-works), i18n planFree1,
admin/docs/ai-gateway.html
- verify-public-claims.py asserts the number AND the period per plan: a right number on the wrong
window is the same defect as a wrong number
newOS4 bundle rebuilt (module-billing), LATEST_VER v4.6.135 -> v4.6.136, main.html cache keys bumped.
Proof:
- verify-public-claims.py rc=0 (copy == policy == live billing API, including periods)
- /api/billing/subscription: limit_usd 0.1, period "day", remaining 0.1
- gateway: gemma4 200, kimi-k2-thinking 502 (gate gone; fake model => upstream 404)
- smoke-auth PASS, smoke-gateway PASS (limit=$0.1 in the payload check)
- live pages serve "$0.10 of AI usage / day", zero $0.50 left on index/pricing/subscriptions
- allowance math: free cap_cost $0.0714/day, pro $7.7779/mo, team $23.35/mo, ent $77.85/mo
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
asking a model (v4.6.135)
`[ -f "/qt.txt" ] && echo exists` β the probe that closed the v4.6.132 batch β was not a shell test
at all: both commands were unregistered, so the line fell through to the AI fallback and a model was
asked whether a file exists. `_evalTest()` has implemented `-f -d -e -z -n ! = !=` for the if/while
parser since the script runner was written; only the command was missing.
* `test` and `[` registered in ShellModule. Success resolves silently; false rejects with a
`silent` error so `&&`/`||` skip it β POSIX prints nothing on exit 1.
* `[` requires the closing bracket (`[ : missing ']'`), and that message is NOT silent.
* `test` already names the on-demand regression-suite runner ("test auto" / "test list" in
help.html). A argv that does not look like a test expression (`-f`/`-d`/`-e`/`-z`/`-n`/`!`,
or `a = b` / `a != b`) is delegated to the handler registered before us, so `test auto`
still runs the suite and the two meanings cannot collide.
* terminal.dispatchCommand() now honours `err.silent` β without it a top-level `test -f /nope`
printed `Error in ShellModule.test:` (measured on v4.6.134, fixed in v4.6.135).
* help.html documents the bracket form next to the existing `test` row.
Three tags were minted minutes apart β v4.6.133 (commands), v4.6.134 (suite delegation),
v4.6.135 (silent top-level) β each a real cache-bust for a fix that superseded the previous build;
the net change is this commit.
Verification
node modules/test/test-shell-test-cmd.js -> 24/24 pass (RED against HEAD: "test command
registered" fails, "m.commands.test is not a function")
node modules/test/test-pipe-guard.js -> all seven flags true
node modules/test/test-parse-args.js -> 28/28 Β· daily job -> node tests 22/22 pass
newos4-maintain.py -> atlas 355 commands (was 354), modules-index 349, help 364/364
live v4.6.135, read from the terminal DOM:
[ -f /qt.txt ] && echo F135 -> F135
test -f /nope || echo M135 -> M135
[ -d /nope ] && echo BAD || echo N134-> N134 (false branch skipped, nothing printed)
test -> no output (suite not loaded; POSIX false)
test -f /nope -> no output
js "a b".length | wc -c -> 1 (v4.6.132 pipe fix still intact)
The AI fallback is no longer reachable for a bracketed file test.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
v4.6.131 taught the pipe stages to forward their raw text, but the shell's own wrapper still split
each stage with `cmdText.split(/\s+/)` β so `grep "alpha beta" /qt.txt | wc -l` handed grep the two
args `"alpha` and `beta"` and matched nothing. Measured live: the wrapper (ShellModule.init() replaces
terminal.executePipeChain) is the implementation that runs, so the core fix alone is decorative.
All six command-line split sites in modules/ext/shell.js now go through one helper, `_tokenize()`,
which delegates to `security._parseArgs` β the parser the bare-command path already uses β so the two
cannot drift again:
111 single command in a `&&` / `;` chain -> `grep "a b" f ; echo done`
176 pipe stage (the wrapper that runs) -> `grep "a b" f | wc -l`
358 `<` redirect input -> `cat < /f`
401 pipe chain inside the script runner -> also forwards the stage's raw text now
424 single command in the script runner -> also forwards raw text
778 `[ -f "path" ]` / `test -f "path"` -> quoted paths in test expressions
Empty stages are skipped instead of being dispatched as a command named '' (`help | | wc -l`).
The guard test grew the assertion that would have caught this: test-pipe-guard.js now checks
`shellTokensQuoted` (tokeniser used, no `cmdText.split(`, empty-stage guard present) alongside
`shellForwardsRaw`; both are RED against HEAD's shell.js.
Verification
node modules/test/test-pipe-guard.js -> {"empty":[],"blank":[],"classifyGuarded":true,
"pipeGuarded":true,"shellForwardsRaw":true,"shellTokensQuoted":true}
node modules/test/test-parse-args.js -> 28/28 Β· daily job -> node tests 21/21 pass
newos4-maintain.py -> maintenance result: OK (index 349, help 363/363, changelog v4.6.132)
live v4.6.132, read from the terminal DOM (no print monkey-patch):
cat /qt.txt -> alpha beta
grep "alpha beta" /qt.txt -> 1 match, 1 file
grep "alpha beta" /qt.txt | wc -l -> 2 (match line + summary; 1 when the quotes are lost)
grep "alpha beta" /qt.txt ; echo done-> match count + done
ShellModule._tokenize(c) === security._parseArgs(c) for 5 sample lines -> true
_tokenize('') / _tokenize(' ') -> [] / []
Also closes the last open feedback item: RPT-20261027-160751 ("Connection failed on all providers")
-> resolve, recorded in decisions.jsonl with evidence (headless probe URL `?probe=ctx`, v4.6.117,
single occurrence, no stack; Ollama answers 200 with 37 models now). Feedback queue: awaiting 0.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
status (v4.6.131)
Four items closed in one batch; three ran as parallel subagents on disjoint files, the bundle was
rebuilt once by the parent (rebuild-bundles.py serialises the working tree, which carries other
sessions' work β see 161372f).
1. PIPE STAGES KEEP THEIR QUOTES. `js "a b".length | wc -c` reached eval as `a b.length` and printed
a ReferenceError. terminal.js executePipeChain() now keeps `stages[i]` aligned 1:1 with
`pipeParts[i]` (the `.filter(argv => argv.length > 0)` shifted the index; replaced by a lastStage
scan + `if (argv.length === 0) continue`) and forwards the stage's own text to the dispatcher.
The core path alone was dead code: ShellModule.init() replaces terminal.executePipeChain
(shell.js:144), and that override is what runs in the browser (`executePipeChain.toString()`
contained `cmdText.split`, not `lastStage`). Both now pass `cmdText`/`pipeParts[i]`.
test-pipe-guard.js was rewritten for the new shape and now also asserts the shell override
forwards raw text β the assertion that would have caught the shadowing. RED on HEAD: both flags
false; GREEN now.
2. MAINTENANCE --fix CAN CLEAR ITS OWN WARN. check_index() ran before check_atlas(), which is what
regenerates course/atlas.json, so --fix rebuilt modules-index.json from the stale atlas and the
same WARN reappeared every run. check_atlas() now runs immediately before check_index().
modules-index.json regenerated: cost/usage present, 349 commands, version 2026-10-29-atlas.
3. `cost` DOCUMENTED. modules/core/module-billing.js registers `cost [on|off]` (an alias of `usage`);
help.html had no row. Added above the `usage` row. help coverage 362/363 -> 363/363.
4. SHIPPED PLANS STOP ASKING FOR A DECISION. The applier only flips review<->safe, so every executed
review-gated plan kept counting as an open decision (5 of 5). newos4-feedback-collect.py now reads
the body: a `## Status` / `## Implementation` heading followed by shipped|executed plus a version,
commit or date is `status: done`; those stay in `plans` (visible as shipped) but leave
plans_review and awaiting_decision. Counts: plans_review 5 -> 0, awaiting_decision 4 -> 1
(the remaining one is a real open report, RPT-20261027-160751).
admin/feedback.html read `it.status[0]`, which would render `d` for a string status; it now
prefers `status_lines[0]`.
Also carried, authored by earlier uncommitted sessions and already live in the bundle (attributed,
not silently absorbed): modules/core/module-billing.js (new, BillingModule: cost/usage),
its CORE_FILES entry in scripts/rebuild-bundles.py, and the BillingModule registry entry in
terminal.js. The bundle was rebuilt from HEAD + this batch with ollama-manager.js taken from HEAD,
so another session's in-flight `_httpError` refactor is deliberately NOT in the shipped bundle
(grep -c _httpError core-bundle.js = 0).
Verification
node modules/test/test-pipe-guard.js -> {"empty":[],"blank":[],"classifyGuarded":true,"pipeGuarded":true,"shellForwardsRaw":true}
node modules/test/test-parse-args.js -> 28/28 pass
maintained job -> node tests 21/21 pass
scripts/newos4-maintain.py -> maintenance result: OK (modules-index 349, help 363/363,
changelog current v4.6.131)
live v4.6.131 (real DOM, no print monkey-patch):
js "a b".length | wc -c -> 1 (js printed 3, one byte)
js "a b".length -> 3
js "abc".toUpperCase() -> ABC
echo "two words" -> two words
node -e "import('fs')" -> red error + Node-builtin hint
js 6*7 -> 42
unhandledrejection -> 0
livePipeImpl: shell-override, livePipeForwardsRaw: true, costRegistered: true
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
command line (v4.6.130)
_parseArgs() treats quote characters as delimiters and drops them, so every string literal in a
js expression died before eval: `js "abc".toUpperCase()` reached the module as
`abc.toUpperCase()` β ReferenceError: abc is not defined. Same for `js import('fs')` β
`import(fs)`. Found while verifying the v4.6.128 console fix
(plan 2026-10-29_112000-newos4-parseargs-strips-js-string-literals.md).
- terminal.js: checkAndRun() records the untokenised line (`_currentRaw`), _dispatch() forwards it,
dispatchCommand(cmd, args, rawText) passes it to the module. Security classification still runs
on the tokenised form, and pipes/AI/script-runner calls pass nothing, so they behave as before.
- console.js: `js` evaluates _expr(args, rawText) β the raw remainder. `node -e` deliberately does
NOT: the tokenizer already keeps `'β¦'` inside `"β¦"`, and the raw form would hand eval the string
literal `"import('fs')"` instead of the import (caught by the live probe, not by review).
js-run keeps tokenised paths.
- modules/test/test-parse-args.js: tokenizer table (records the deletion as the current contract),
dispatcher plumbing asserts, _expr cases, and the user-visible paths. 27 assertions.
- module-loader.js + main.html + bundle: v4.6.130 (129 was minted, then console.js changed again β
re-minting so no cached client keeps the intermediate).
Verification
node modules/test/test-parse-args.js -> 27/27 pass
same file against HEAD sources (git show HEAD:...) -> TypeError: _expr is not a function (RED)
scripts/lint-no-undef.py terminal.js console.js -> clean, 122 cross-file globals
core-bundle.js rebuilt from HEAD sources + this patch in /tmp/iso (the working tree carries
another session's uncommitted ollama-manager.js `_httpError`); diff vs HEAD bundle = 14 lines,
all mine. critical-ext-bundle.js byte-identical.
Live page v4.6.130 (real executeCommand, output captured):
js "abc".toUpperCase() -> ABC
js 'a b'.split(' ').length -> 2
js JSON.stringify({a:1}) -> {"a":1}
js 6*7 -> 42
js import('fs') -> red "Failed to resolve module specifier 'fs'" + Node-builtin hint
node -e "import('fs')" -> same error + hint (was printing the expression as a string)
echo "two words" -> two words
js-run /nope.js -> js-run: /nope.js: No such file
unhandledrejection count -> 0
The sibling's parked terminal.js hunk (BillingModule registry entry) was stashed for this commit and
is restored in the working tree afterwards; it is not part of this change.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
(v4.6.128)
evalJS() evaluated the expression bare and printed JSON.stringify(result), so a
rejected thenable (e.g. `js import('fs')`) became `{}` on screen while the rejection
escaped to window.onunhandledrejection and was auto-filed as a runtime bug report
(RPT-20261028-172211-9c7f2adb, cluster 480x). evalJSFile() already had .then/.catch β
the asymmetry was the defect.
- modules/ext/console.js: thenable branch in evalJS() awaits and catches; new
_printError() is the single error sink for all four catch paths and appends a hint
when a bare specifier fails to resolve ("'fs' is a Node built-in; the browser has no
Node β use the VFS (ls/cat/run, terminal.fs) instead").
- help.html: js / js-run rows say "browser only β no Node built-ins".
- module-loader.js + main.html: v4.6.128.
Verification (/tmp/test-console-fs.js, 11 asserts, new Function harness over the module):
GREEN 11/11 β rejection prints one red error + hint, no unhandledRejection; resolved
thenable prints its value; sync value/throw unchanged; js-run file + missing file OK.
RED 4/11 against .bkp-console-20261029_121243/console.js β A1/A2/A3 print `{}`,
A4/G2 leak "Failed to resolve module specifier 'fs'".
node --check modules/ext/console.js β OK.
console.js is loaded separately via extKeys (not in critical-ext-bundle.js), so no bundle
rebuild β proven by `grep -c JSConsoleModule modules/ext/critical-ext-bundle.js` = 0.
Consumes STEER-20261028-195043-8084. Plan: .hermes/plans/2026-10-29_063000-newos4-js-console-fs-specifier.md
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Found by running the shipped runner the way a user would, not by re-reading it.
- _evalTest() stripped [ ] but not the `test` keyword, so `if test -f /work/alpha.txt` evaluated
the word "test" as an operator, returned false, and every such branch took the else arm β
measured live: "alpha missing" for a file that exists. Both spellings work now
(`test -f x`, `[ -f x ]`), plus `-d`, `-e`, `-z`, `-n`, `!` and the = / != comparisons.
- `for f in /work/*.txt` ran ONCE with $f set to the literal pattern, because only command
arguments went through _expandGlobs. Loop items are expanded per iteration now, so the body runs
once per matched file with the real path.
- _expandGlobs() falls back to readDir() when a fs has no ls(), so the helper works against either
shape instead of silently returning the literal pattern.
Verification, live on v4.6.119:
- for+glob: "--- /work/alpha.txt / 3 /work/alpha.txt / --- /work/beta.txt / 1 /work/beta.txt"
(6 statements, exit 0) β previously one iteration printing the pattern.
- conditionals: "alpha exists", "nope missing", grep + $? intact.
- end-to-end /work/report.sh (loop + glob + >> append + test -f) wrote a correct log:
"start / file: /work/alpha.txt / 3 /work/alpha.txt / file: /work/beta.txt / 1 /work/beta.txt β¦".
- unknown command still reports "command not found" and never calls the AI.
- test-shell-script.js 27 -> 35 asserts.
Retracted: I reported `echo x > file` as broken. It was my probe monkey-patching terminal.print,
which bypassed the capture stack the redirect uses. Both paths write correctly:
direct helper -> "direct-call", prompt -> "prompt-call\n".
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
as one line
Two things I had flagged as "worth your attention" and one leftover nit.
- num_ctx was never sent. Ollama's server-side default context is 2k-4k depending on the build, so
a system prompt + 20 messages + up to 24k chars of tool results could be truncated server-side,
which is exactly the "it forgot what it just read" class of report. Every Ollama request now
carries options.num_ctx (default 16384, overridable via webx11_ollama_numctx, floor 512); the
OpenAI-compatible branch is untouched because num_ctx is not one of its parameters.
- Automatic connect prints one outcome instead of one line per endpoint tried (7 -> 3 on a dead
endpoint). Interactive `ollamaconnect` keeps the per-endpoint detail.
- write_file's failure message no longer echoes a trailing slash.
Verification:
- test-ollama-numctx.js (7 asserts): generate and chat payloads carry num_ctx=16384, a stored
override is used, junk/too-small values fall back, and the /v1/chat/completions branch does not
receive it. RED on HEAD: TypeError reading 'num_ctx' of undefined. Suite 20/20.
- Live v4.6.117, capturing the real request: {"num_ctx":16384} on /api/generate with
model gemma4:31b-cloud, reply "OK".
- Live v4.6.118, dead endpoint (127.0.0.1:9, fallback disabled): automatic -> 3 lines
("connecting on first useβ¦", "Connecting to deadβ¦", "Auto-connect failed β β¦"); manual
connect() -> 4 lines including "endpoint http://127.0.0.1:9 failed", so the suppression is
conditional, not blanket.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The stdin work in 5cf5010 fixed wc/head/tail but left grep alone, and grep's no-operand branch
sweeps the cwd. Measured on v4.6.115: `cat /gdir/in.txt | grep zzz` answered
`other.txt:1: zzz in a cwd file` β a file that was never piped, and `cat in.txt | grep bbb`
reported the match twice (`in.txt` from the cwd sweep plus `__pipebuf__` from the injected file).
- module-fs.js: when terminal._pipeInput is set and no file operand was given, grep searches the
staged text and prints `<line>: <text>` (stdin has no filename, as in GNU). The cwd sweep stays
for a bare interactive `grep x`.
- shell.js: _isPipeAware now covers only sort/cut/uniq/expand. grep/head/tail/wc read _pipeInput
themselves, so injecting the __pipebuf__ file produced duplicate hits and named stdin results.
Verification:
- test-shell-stdin.js 20 asserts (+4 for grep: no cwd match, non-match reports none, match prints
no filename, -c counts stdin). RED on HEAD: G1/G2/G3 fail with `other.txt:1: zzz in a cwd file`.
- Suite 18/18 files; maintain --fix "node tests 18/18 pass".
- Live v4.6.116 after `cd /gdir` (where other.txt contains zzz):
`cat /gdir/in.txt | grep zzz` -> (no matches) Β· `| grep bbb` -> `2: bbb` Β· `| grep -c b` -> 1 Β·
`| head -n 2` -> aaa,bbb Β· `| tail -n 1` -> ccc Β· `| wc -l` -> 3.
Credit: the false-positive grep was reported by the stdin subagent as an aside; re-derived here
against the live terminal before fixing.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
globs, redirects, stdin
MAINT-021 and MAINT-017, executed from .hermes/plans/2026-10-27_145500-newos4-script-runner.md.
- shell.js: new block parser (_parseBlocks) and executor (_runInstructions/_runSimple/_runCommand/
_dispatchOne). `for`/`while`/`if` bodies now run, including one-line `if [ -f x ]; then y; fi`.
Variables expand through _expandVars ($VAR, ${VAR}, $?, single quotes literal), globs expand
through _expandGlobs (unquoted args only, per-command), and `>`/`>>` capture the command's output
into the VFS instead of being passed to the command as text.
- shell.js: a script never answers an unknown command by calling the AI any more. It prints
`script:<path>:<line>: <cmd>: command not found` and sets $? β the AI fallback stays for typing.
- shell.js: unclosed blocks and bad `for` headers surface as `Script <path> failed: β¦` (the parse
used to sit outside the try, so a bad script crashed the runner).
- shell.js: interactive path gets the same treatment β prompt globbing and `echo hi > /f` now work.
Cause of the asymmetry: the two code paths had separate dispatch.
- shell.js: `_isPipeAware` no longer injects the __pipebuf__ file into wc (it reads _pipeInput), and
the module registers itself with a versioned URL and calls its own init() β registerModule() does
not call init(), so the executeCommand wrapper was never installed from the boot bundle.
- module-fs.js: wc/head/tail/grep read piped stdin when no file operand is given; `cat` accepts
several operands (`cat *.txt`); `ls <file>` names the file instead of printing "(empty)".
- main.html: shell.js removed from the extension list β it is already inside
critical-ext-bundle.js, and loading it twice produced two ShellModule instances that could
disagree about their version.
Verification:
- test-shell-script.js 27 asserts (parser, variables, globs, executor, redirect, command-not-found,
one-line if, parse error). RED on HEAD: `TypeError: shell._parseBlocks is not a function`.
- test-shell-stdin.js 16 asserts (added by a subagent): wc -l/-c, head -n, tail -n over stdin, and
the file-argument path unchanged.
- Suite 18/18 files; maintain --fix "node tests 18/18 pass".
- Live on v4.6.115, through the real executeCommand:
`ls /n*.txt` -> /n1.txt Β· `cat /n*.txt` -> one|two Β· `echo hi > /r.txt` then `cat /r.txt` -> hi Β·
`cat /probe.txt | wc -l` -> 3 (bare, no __pipebuf__) Β·
`script /run.sh` -> item 1 / item 2 / item 3 / exists / command not found, exit 1, no AI routing.
|
| |
|
|
|
|
|
|
|
|
|
| |
MAINT-004. The guard read warnAt/blockAt/hardStopAt while one config path wrote `threshold`, and
`enabled` was never consulted anywhere β writing {"enabled": false} into webx11_ai_config changed
nothing at all. One resolver, _loopPolicy(), now normalises both spellings, guarantees
warnAt <= blockAt < hardStopAt, and both the block and the warning are gated on enabled.
Live on v4.6.110 (_loopPolicy called in the page): defaults {warn 3, block 6, hardstop 9};
{threshold: 4} -> {warn 2, block 4, hardstop 7}; {enabled: false, threshold: 3} -> enabled false.
test-loop-policy.js: 17 asserts, RED on HEAD ("_loopPolicy is not a function"), suite 16/16.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Three of the four agent-cluster items, each with a test that is red on HEAD.
- MAINT-008 (module-agent-registry.js + module-orchestrator.js): registerAgent took args[0] as the
agent id and only args[1] as the capability list, while both orchestrator callers passed the
command word first β so the boot log read "Agent 'register' registered with 1 capabilities" and
four capabilities were stored as one. The command word is now stripped, args[1..] are joined and
split on commas (both "a,b" and "a b c" work), and the callers pass [id, ...caps].
- MAINT-007 (registry + aiassist): mkdir/rm used relative paths, so a registration made from /docs
created /docs/agents/<id> and an unregister from any other cwd removed nothing β the leaked
directories the report suspected. Every path in the registry (and the four /_sessions writes in
module-aiassist.js) is absolute now.
- MAINT-009 (listAgents): a directory alone counted as a registered agent, so a hand-made
/agents/foo appeared as "foo β unknown (0 capabilities)". Only a readable status.json with an id
counts now; the rest are reported as unregistered directories.
Verification: test-agent-registry.js (11 asserts). RED on HEAD: 9 of 11 fail, including the exact
`β
Agent 'register' registered with 1 capabilities` line and `hand-made β unknown (0
capabilities)`. Suite 15/15 files green; maintain --fix "node tests 15/15 pass".
Live on v4.6.109: boot prints "β
Agent 'orchestrator-0' registered with 4 capabilities";
`agent-list` warns "1 directory(ies) in /agents have no status.json (not registered): hand-made";
register from / with a comma list stores 3; unregister from /docs removes the directory and leaves
no /docs/agents.
MAINT-006 stays open with a narrower note: the hard stop was only reachable when every call in a
batch was blocked, so a mixed batch (one repeated + one novel call) never ended. The same ceiling
now applies on the mixed path, but that branch lives inside agenticLoop and is not covered by a unit
test β it needs a live agent run that repeats a call.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
leave /skills
Two board items plus the context-window close-out.
- MAINT-019 (module-system.js): `help <cmd>` answered "No help available for X" for a command
that is registered but has no help line, which reads as if the command does not exist. It now
says "X β registered by <Module>, which publishes no help line for it. Try: help". A command
that genuinely does not exist still gets the old message.
- MAINT-020 (module-aiassist.js): auto-generated loop lessons were written into /skills with
their own index entries, so `skill list` filled up with system noise. Lessons now live in
/lessons with /lessons/index.json; the reader still falls back to legacy loop-lesson-* entries
in /skills for one release, and the one-time migration deletes them from the skills index.
Verification: test-help-and-lessons.js (10 asserts, includes a lint-style guard that no lesson is
written into /skills). Suite 14/14 files green; maintain --fix "node tests 14/14 pass",
"help coverage 362/362". Live on v4.6.108: `help agent-register` -> "registered by
AgentRegistryModule, which publishes no help line for it"; `help a2a-prune` unchanged;
`help nope-not-real` still "No help available".
MAINT-001 close-out (no code change here): history is capped at 20 messages with a "[context trim]"
marker (first trim at turn 11 in a 60-turn probe), token compression took 80 msgs/8080 tok to
17/2625, and since v4.6.107 one turn's tool bodies are clipped to 6 000 chars each and elided past
24 000 with the newest 4 kept verbatim. Residual, recorded on the board: maxContextTokens is never
sent to the server as num_ctx, so the real window is whatever the endpoint defaults to.
MAINT-018 close-out: not reproducible. Live probes on v4.6.107/108 β `sed -n '1,3p' /probe.txt`
prints lines 1-3, `sed -n "2,4p"` prints 2-4, `grep 'line4'` matches, and the same commands inside
`script /q.sh` print correctly. The "Command not found: for;" symptom belongs to the line-by-line
script runner, which is MAINT-021 and stays open.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
a corrupt memory graph
The three items the last review and probes left open, each with a test that is red on HEAD.
- MAINT-023 (module-aiassist.js): one agentic turn grew without limit β every role:'tool' body
was pushed verbatim and only this.context was trimmed BETWEEN turns (measured 6 -> 112 705
tokens across 8 model calls against maxContextTokens 12 000). Observations are now clipped to
maxToolResultChars (6 000) with a note saying how much was withheld and to re-read with
offset/limit, and _boundToolContext() elides the oldest tool bodies once the running total
passes toolContextBudgetChars (24 000), always keeping the newest 4 verbatim. The ceiling is
turn-independent: the same ~26 k chars at 8 turns and at 20.
- MAINT-024 (modules/ext/skill-pack-core.js): the command printed "β
Skill created" with
fs.writeFile()'s boolean discarded (probe: fresh VFS, success printed, no file), stripped
invalid characters instead of rejecting the name (My-Skill -> y-kill, DevOps_Thing ->
evpshing), and treated an unreadable /path argument as the skill body. Now: kebab-case is
validated (uppercase normalises and the final name is reported; anything else is refused),
/skills is created on first use, the write and index results are checked, and an unreadable
body path is an error.
- MAINT-003 (module-memory-graph.js): a corrupt store looked exactly like a fresh install β
fsReadJSON returns null on a parse failure, so 85 nodes could vanish with no warning and no
copy. The probe confirmed the loss: seed 6 nodes, truncate nodes.json and export.json, reload
-> 1 node, no message, nothing in .trash. _quarantineCorrupt() now keeps the unreadable file
as <name>.corrupt, warns, and then loads from whatever remains.
Tests: test-tool-context-bound.js (11), test-skill-create.js (16), test-memory-graph-recovery.js
(8). RED on HEAD: tool-context crashes on _clipToolResult (exit 1), skill-create 16 assertions
fail on the missing file/index, memory-graph B1/B2/C1 fail (no warning, no .corrupt copy).
Suite now 13/13 files green; maintain --fix reports "node tests 13/13 pass".
Live on v4.6.107: skill-create My-Skill -> β
"my-skill"; DevOps_Thing -> Invalid skill name;
/does/not/exist.md -> Cannot read skill body; only my-skill in the index.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
review findings
Second pass over v4.6.105, driven by an independent review that reproduced each item.
- ollama-manager.js: chat() now reports the captured cause ("Ollama error: Failed to fetch",
"β¦: HTTP 500") instead of the generic "No provider responded"; _lastChatError is reset per
call so an old reason is never replayed. chatWithTools carries the reason in its error string.
- ollama-manager.js: _ensureConnected() bounds connect() with a deadline (connectTimeoutMs,
default 10 s). connect() uses a bare fetch(), so a blackholed endpoint previously left the
shared in-flight promise pending forever and hung every later message β the old code failed
fast. The attempt is now try/finally, so a throwing terminal.log() can no longer poison
_connecting and suppress the retry on the next message.
- ollama-manager.js: an automatic connect no longer silently rewrites the user's saved model;
the previous choice is restored in memory and in localStorage, and the user is told.
- ollama-manager.js: the DeepSeek empty-content branch returned top-level tool_calls with
content '', so the caller pushed an assistant message with no tool_calls and then a
role:'tool' reply β an orphan tool message, which is the HTTP 400 that branch exists to
prevent. tool_calls now live inside the message as well.
- module-aiassist.js: a no-argument list_files returned the ROOT instead of the cwd (readDir('/')
instead of ls()); it now means the current directory again. Error text keeps the path as the
caller wrote it, and write_file distinguishes a read-only virtual path (/sys/β¦) from a missing
parent instead of blaming the parent.
Tests: test-ollama-error-text.js (new, 14 asserts, RED 9/5 before the edit), test-ai-fs-tools.js
7 β 16 asserts (write onto a directory, /sys write, list a file, '/', no-arg cwd, relative path),
test-ollama-autoconnect.js 11 β 19 (blackholed endpoint settles + retries, saved model survives;
the old "connecting hint" assertion is now falsifiable β the reviewer showed it also matched
connect()'s pre-existing line). Suite: 10/10 files green. Live on v4.6.106: healthy auto-connect
42 ms with the hint; blackholed endpoint settles in 1242 ms with connectTimeoutMs=1200, prints
"timed out" and "unreachable", _connecting cleared.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
sees /sys
Three defects, each re-probed first and covered by a test that is red on HEAD.
- ollama-manager.js: nothing connects at boot; only the Ollama tab button called connect(),
so a user's first message died with "Ollama not connected. Use ollamaconnect first." even
when the endpoint was healthy. New _ensureConnected() runs from chat(), chatStream() and
chatWithTools(): one attempt, shared by concurrent callers, a "Connectingβ¦" hint, and only
then the error β which now says the endpoint is unreachable instead of telling the user to
run a command they already expected to work.
- module-aiassist.js write_file tool: fs.writeFile()'s false return was dropped and the tool
answered "File written: β¦" anyway, sending the model on to a read that found nothing
(MAINT-012). It now returns "Write FAILED: β¦ β parent directory missing or target is a
directory; nothing was written".
- module-aiassist.js list_files tool: the hand-rolled walk over root.content never resolved
the virtual mounts, so /sys β a directory the UI and `ls` both show β answered
"Directory not found" (MAINT-011). It now uses fs.readDir(), which resolves mounts and
returns null for a genuinely missing path.
Tests: modules/test/test-ai-fs-tools.js (7 asserts) and
modules/test/test-ollama-autoconnect.js (11 asserts). RED on HEAD: ai tools 4 passed/3 FAILED
(/sys β "Directory not found", write_file β "File written: β¦"), ollama 2 passed/9 FAILED
("Use ollamaconnect first"). GREEN now 7/7 and 11/11; maintain --fix reports "node tests 9/9 pass".
Live on v4.6.105: connectedBefore=false, _ensureConnected()=true in 41 ms, 37 models.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
MAINT-010 and MAINT-016, the two high items the review left open. Both were re-probed
before the change and the new test is red on HEAD (5 passed, 7 FAILED).
- virtual-fs.js rename()/copy(): they assigned dst.dir.content[dst.name] with no destination
type check, so `mv file /docs` replaced the /docs node and every child vanished while the
command printed "Moved: β¦". Refuse when the destination is a real directory node (the
virtual /sys entry is exempt). rename() also erased the file on `mv a a`, because assigning
and then deleting the same key removes it; that path is now an explicit no-op.
- module-fs.js mv/cp: a refusal used to surface as "No such file or directory", which sent
the operator looking for a missing source. The destination is now inspected first and the
message says "is a directory β refusing to overwrite it".
- module-fs.js ls: the path was only read from argv[0], so `ls -l /work` listed the cwd.
The first non-flag argument is now the target; the path-first form is unchanged.
Tests: modules/test/test-fs-mv-cp-guard.js (12 asserts) β RED on HEAD (A1-A3, B1, B2, C2, E1),
GREEN now; maintain --fix reports "node tests 7/7 pass". Browser on v4.6.104:
"mv: /vtest: is a directory β refusing to overwrite it", child and source intact,
"ls -l /vtest" lists /vtest (keep.txt), not the cwd.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
gate the build
Review of the 08:30 implementer run (v4.6.102, checkpoint name fix). The shipped fix was
correct; these are the defects it introduced or left behind, each measured first.
- virtual-fs.js rollback(): the snapshot strips _snapshots/_sessions/.trash for bloat. Restoring
it verbatim deleted every checkpoint, every agent session and the trash, made rollback
irreversible and unrepeatable. Re-attach the live copies after the root swap.
- checkpoint(): names are millisecond-resolution, so tight calls overwrote each other
(measured 6 calls -> 3 snapshots). Dedupe with a -N suffix, refuse names containing "/"
(written but unlistable), and prune by snapshot createdAt instead of filename sort, which
let a custom name ("aaa-oldest") evict newer auto-named checkpoints.
- module-predictor.js: legacy pre-v2 points have no `usable`, so a 0.079 s aborted run
re-entered similarity as evidence and floored every prediction at 1 s. Backfill `usable`
from the measurement on load, drop the ones it rejects, add a belt-and-braces guard in
_findSimilar and count only usable points as actual executions.
- newos4-maintain.py: run every node-runnable modules/test/test-*.js on each maintenance pass.
The v4.6.102 regression was written but no job ever executed it.
- newos4-steer.py: `done` now requires --commit <sha> that exists in the repo, and gains
`reopen`. A written plan was being closed as done at 06:17 while the predictor fix shipped
nowhere, so the implementer's pre-flight queue was empty.
- steering.json: predictor item reopened, checkpoint item stamped with f7249ff.
- maintenance board: MAINT-002/005/015 were fixed in v4.6.81-85 but still listed open/high;
closed with their commit ids, and MAINT-010/016 annotated with the probes that keep them open.
Tests: test-vfs-checkpoint 23/23 (RED 14/9 before), test-predictor-accuracy 38/38
(T1-T3 RED on HEAD), maintain --fix all PASS, browser round-trip
"Checkpoint created: <name>" -> "Rolled back to checkpoint: <name>" on v4.6.103.
Also carries the already-deployed 2026-10-26 syntax-gate hunk in newos4-maintain.py (syntax_failures + gated rebuild) that had never been committed.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Task 2 + ship of plan 2026-10-27_061724-newos4-checkpoint-name-mismatch.md.
module-system.js `checkpoint` read the clock a second time for the printed
name, so the name in "β
Checkpoint created: <name>" differed from the name on
disk by >=1ms whenever no explicit name was given. It now prints the value
returned by fs.checkpoint(). try/catch and error strings unchanged.
Bundles rebuilt (rebuild-bundles.py); LATEST_VER in module-loader.js and all
three ?v= tags in main.html (core-bundle, module-loader, critical-ext-bundle)
bumped v4.6.101 -> v4.6.102.
Verification:
$ node --check modules/core/module-system.js
SYNTAX OK
$ node --check modules/core/core-bundle.js
BUNDLE SYNTAX OK
$ node modules/test/test-vfs-checkpoint.js (x5)
test-vfs-checkpoint: 10 passed, 0 failed
$ curl -s https://webx11.duckdns.org/spaces/newos4/main.html | grep -o 'core-bundle.js?v=[^"]*'
core-bundle.js?v=v4.6.102
$ python3 scripts/newos4-maintain.py
PASS: version v4.6.102 x3 in main.html (expect 3) ... maintenance result: OK
$ python3 scripts/newos4-skill-audit.py
skill audit: 7/7 pass
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Task 1 of plan 2026-10-27_061724-newos4-checkpoint-name-mismatch.md.
virtual-fs.js checkpoint(name) generated an ISO name when none was given but
returned bare `true`, so callers could not know which name the snapshot was
written under. It now returns the resolved name (string). Prune (max 3) and
saveToStorage() behaviour unchanged.
Verification:
$ node --check modules/core/virtual-fs.js
SYNTAX OK
$ node modules/test/test-vfs-checkpoint.js
test-vfs-checkpoint: 9 passed, 1 FAILED (remaining FAIL was the
module-system display path, fixed in the next commit)
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The two panes kept separate clipboards (filemgr._clipboard vs webdav._selected), so "cut here, paste
there" did not exist β the operator's last phone-test item.
filemgr.js _openOverlay() publishes a read-only view of its clipboard on the terminal object:
webx11.fileClipboard() / fileClipboardClear() / fileClipboardSet(paths, mode).
webdav.js the bulk bar gains "π₯ Paste N" (hidden when the clipboard is empty). Handler: for each
path readFile -> _webdavRequest('PUT', dest + '/' + name, {body}); only 200/201/204 counts.
For mode "cut" the local file is deleted AFTER the server accepted the bytes, never
before. A failed PUT keeps the local copy, prints "π₯ Paste failed: <name> β HTTP n",
and leaves that path in the clipboard so "π₯ Paste 1" is a retry. Successes leave it.
Summary "π₯ Pasted N item(s) to <path>" (+ "β M failed"), then the pane reopens there.
Semantics: cut moves (upload then delete local), copy leaves the local file and keeps the clipboard so
the same files can go into several folders.
Verified with the transport stubbed in the live page (this browser has no WebDAV credential):
button "π₯ Paste 2" for a 2-path cut
PUT ok (201) -> local file null (moved)
PUT fail(500) -> local file "BBB" (kept)
clipboard after -> {mode:"cut", paths:["/fail-d.txt"]}, button "π₯ Paste 1"
node --check both modules, no-undef clean (55 files), v4.6.101 x3 in main.html, syntax OK, help 362/362.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
share, dead-end fullscreen, silent WebDAV move
All four came from the operator's phone pass on v4.6.95 and each collapsed to a line of code.
A "No file selected to download": the preview pane had no close control; the only β was the file
manager's own, and it calls _closeOverlay() which wipes _multiSelect/_selectedFile/_selectMode.
On a phone the preview covers the list, so closing it before pressing Download always emptied the
selection. The toolbar now starts with β -> _clearPreview(), which keeps the selection and leaves
fullscreen first; Download with no targets names the fix instead of stating a dead end.
B share/copy: π existed only in the preview toolbar and took a single path, and after the Android
share sheet there was no copy affordance. Added _sharePaths(paths, mode) used by both the toolbar
(1 path, π copy and π share) and a new action-bar π Share (whole selection), plus
_inlineLinkBox() which renders a selectable input/textarea with all URLs when neither the share
sheet nor the clipboard is available. No prompt() anywhere on this path.
C fullscreen: requestFullscreen targeted .fm-preview-body, which sits INSIDE the pane but excludes
the toolbar, so βΆ and β disappeared and only the OS back gesture got out. Now fullscreens
#filemgr-preview β toolbar, β and βΆ stay on screen and βΆ toggles.
D webdav move: bulk Move asked for the destination with prompt(), which returns null in iOS
standalone PWAs and most Android WebViews, so it silently did nothing and looked unimplemented.
Destination is now an inline row in the bulk bar (44px input + "Move here"/"Cancel", prefilled).
Verified on the served code: ext bundle md5 5f440c77b6ab5529729aee8019a6c351 (served == local),
webdav.js carries wdav-move-dest x2, and a headless browser probe on the real page shows the toolbar
as ["β","βΆ","π","π"], closing the preview prints "Preview closed β selection kept" with the
selection intact, selecting 3 files and copying yields a textarea with the 3 real WebDAV URLs, and
Download with nothing selected explains what to do. 6 PASS on the URL test, no-undef clean,
v4.6.99, help 362/362.
Not done, deliberately: cut/copy + paste across the VFS/WebDAV border. Recorded as Issue D2 in the
plan with the exact insertion points, the upload-then-delete ordering rule and the failure case.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
multi-select, Web Share, 44px targets
Executes .hermes/plans/2026-10-24_051600-newos4-mobile-download-share.md (approved via the feedback
dashboard; marker autonomy: safe). Front-end only, v4.6.94 -> v4.6.95.
T1 modules/ext/filemgr.js β URL.revokeObjectURL ran in the same tick as a.click(); iOS Safari and
Android Chrome hand the blob: URL to the download manager asynchronously, so the transfer was
aborted while the status line still said "Downloaded". Now revoked after 30 s.
T2 download honours the multi-selection via _getSelectedPaths() (falling back to _selectedFile),
reports "β¬ Downloaded N files" and "β Skipped <path>" instead of a silent partial success.
T3 share: navigator.share -> clipboard.writeText -> inline read-only <input> (focused, selected).
prompt() is suppressed in iOS standalone PWAs and most Android WebViews, which is why tapping π
did nothing. Unresolvable space now reports "π Share unavailable β no space resolved".
T4 both preview toolbar buttons: min-height/min-width 44px, touch-action manipulation, 1em, gap 10px
(was 24px/0.85em/6px β the smallest tap targets in the UI).
T5 _previewShareUrl falls back _resolveSpaceName -> _detectSpaceName (webdav.js:22) -> page path,
and returns null rather than emitting /spaces//<path>.
T6 the URL helpers are now module-scope pure functions exported for Node, so
modules/test/test-download-url.js tests the shipped code instead of a local copy (4 -> 6 cases).
Also fixes a latent bug that made require() impossible: `typeof window.webx11` throws under Node
because typeof does not protect a member access on an undefined global.
Verified on the served bundle (filemgr.js is concatenated into critical-ext-bundle.js, NOT
core-bundle.js β the plan's T7 grep pointed at the wrong file): HTTP 200, 129 867 B, 2x
navigator.share, 3x fmBuildShareUrl, deferred revoke present, 11x min-height:44px.
Live browser probe on /spaces/newos4/main.html: _previewShareUrl('docs/a.md') ->
https://webx11.duckdns.org/spaces/newos4/docs/a.md (no double slash); with navigator.share absent
and clipboard blocked, clicking π produced .fm-share-input containing that URL; button style ends
min-height:44px;min-width:44px;touch-action:manipulation; toolbar gap:10px.
node modules/test/test-download-url.js -> 6 PASS. lint-no-undef clean (55 files).
maintain: PASS version v4.6.95 x3, syntax OK, HTTP 200 /main.html.
Also adds admin/docs/session-overview-2026-10-26.md (glm-5.3-compressed session record).
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Dashboard (/admin/feedback.html, admin-gated)
- User-feedback tab over the live :8908 API: list, filter, search, publish reply,
archive/reopen. Reply is what the reporter sees in /stear -> Your reports & replies.
- Agent-items tab: 3 plans parked at autonomy:review + 19 defect reports, one click each.
- Snapshot age line + stale warning (>30 min), and the last apply outcome with failures.
- Existing /admin/ tab relabelled "Client error log" (it renders feedback_log.json, not
the feedback store) and now links to the review dashboard.
Decision path
- admin/api/feedback-decide.php: exact-match nginx location + auth_request, Origin check,
item_id allowlist, append-only decisions.jsonl with fwrite+fsync verification.
- scripts/newos4-feedback-apply.py (darkbot): last decision per item wins, superseded rows
recorded, reject reverts safe->review, per-item try/except, ledger saved per mutation,
realpath containment, exact standalone-line marker flip, file mode preserved, queue
rotation at 1 MB. Steering directives now carry the report title.
- scripts/newos4-feedback-collect.py: snapshot for the dashboard, 40-line plan head,
steering shape assert (-1 = unknown), recent_apply outcomes.
- newos4-feedback-sync.sh: flock -n so overlapping runs cannot double-apply.
Guide + terminal
- feedback-guide.html: which channel (terminal feedback / /stear / automatic reports),
expectations, exact payload per channel, how to write a report that gets fixed.
Linked from /stear, index.html footer and help.html.
- newOS4: feedback list|read|reply|archive|reopen|stats|dashboard|help (+ feedback-admin
alias), admin-token derived per call, argument-validated dispatch so a sentence like
"feedback list is broken on mobile" is still a message, "feedback -- <text>" forces it.
- help.html: 8 rows, coverage 362/362.
Review: glm-5.3 REWORK -> 15 findings fixed, 5 refuted with evidence, 2 doc-only.
Guide corrected (invented "60 MB per batch" replaced by the real 512 KB/5-per-10-min limits).
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
(v4.6.91)
P0-5 asked what to do about a 422k-token session against a 400k threshold. There is no 400k
threshold in this tree β the budgets are maxContextTokens (12000) and maxContext*2 messages β so
rather than chase a number that does not exist, this applies the decision that is defensible
either way: warn before the cliff, never lose history without saying so.
- 80% of the token budget prints one warning per crossing (no per-turn nagging)
- the message-count trim unshifts a marker naming how many messages left the window and quoting
the earliest one, instead of splicing them away silently
- 'max-context-tokens (default: 2000)' now reports the real code default, 12000
Verified: 900/1000 tokens -> warns once, second call silent; 1500/1000 -> compresses to 17
messages with the summary head intact; 9-message window with maxContext=2 -> 4 messages, marker
'[context trim: 6 earlier message(s) ... earliest was: "turn 0 ..."]'.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
(v4.6.88-90)
One phantom entry, three symptoms:
- ls() injected 'sys' whenever it was called without a path, so 'cd _agents; ls' advertised a
sys directory that does not exist, and list_files('_agents/sys') answered 'Directory not
found' β the reported false negative. Injection is now gated on dir === root.
- tree() pushed the same virtual name into the walk and dereferenced it, throwing mid-output on
every root listing. Guarded.
- _gatherFileInfo() walked into /sys and counted a synthetic subtree: 146 files / 114 dirs for a
44-file / 15-dir store. Virtual entries are skipped.
Verified: root still lists sys; _agents and skills do not; tree('') completes; world model total
(44) equals an independent walk of fs.root (44), agree: true.
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
(v4.6.86-87)
The reference checker ran /skill-view \S+/ over every skill's prose, so 'skill-view <name>' and a
quoted example in skills-system-guide were counted as missing skills: 'Broken refs: 2' with no
names, both phantom. Placeholders and quoted names are now ignored (trailing punctuation stripped
first), and the survivors are listed in the audit line and in the suggestion.
Verified: clean tree -> Broken refs: 0. A probe skill referencing 'ghost-skill' -> Broken refs: 1
(probe-broken-ref β ghost-skill), while '<name>.' and '"module-management"' in the same file are
ignored.
|
| |
|
|
|
|
|
|
|
|
|
|
| |
getStaleSkills() returned daysUnused=999 for a skill that had useCount 0 and no lastUsed β a
placeholder rendered by the audit as '999d unused', indistinguishable from a measurement. It told
the operator to archive procedural memory that had simply never been instrumented.
Now: daysUnused null + neverUsed true when there is no lastUsed; real day count otherwise; the
audit prints '(never used)'.
Verified: probe skill with no telemetry -> {daysUnused: null, neverUsed: true}; one last used
2026-08-01 -> 83; no 999 anywhere in getStaleSkills() or the audit output.
|
