The token card's title row gains the session's cache-hit share in its
right corner — the harness chat line's own reads-over-billed formula,
reusing the stats board's label and tooltip — hidden until something
was billed. The timing card's throughput chip drops its pill chrome and
bold weight, now reading 输出速度 N tok/s (zh) / TPS N tok/s (en) with
the TPS expansion on the tooltip; both corners share a new plain-text
.lc-card-rate style and keep their native-title tooltips.
The day strip spanned the card's full width while the columns sit inset
between the two 48px tick rails, so every label fanned out from its bar
group; the topmost tick also rode the title's 4px margin. The plot is now
a 3x2 grid — rail | chart | rail, with the strip as row 2 of the chart's
own column, so labels center under their columns by construction — and
the title falls back to the shared 10px margin. Under a 400px container
the seven MM-DD labels thin to every other one instead of crushing.
Fires on issue creation; currently the only blocked pattern is
'dsh-plugin.org' (case-insensitive). Matching issues are closed as
not planned with an explanatory comment. Add further conditions by
appending to the patterns array.
The donut cards' slice legend auto-fills at a 150px floor (from 160px)
with an 8px column gap (from 20px) and a 30px share column (from 34px),
so two columns start at a 308px legend — the Insights page's ~440px
stats cards ride this — and wider cards take more; the floor is
`min(150px, 100%)` so a narrower legend shrinks the track instead of
clipping the share column (the 390px fold previously cut 11px off the
percentages). Narrow panes keep their single column; the Context tab
keeps the same column count at every width it already wrapped at.
The Context Insights surface moves from a sidebar-foot button opening a
frame-wide overlay to a first-level panel, a sibling of the shipped
Plugins and Automation tasks pages: the page registers on the layout's
keyed `main` slot and the entry on `sidebar.panellist` under one id
(`dsh-context`), so the shell owns the entry row and panel switching —
mounting the page IS opening it (no plugin-side open state, the footer
entry, the overlay seat, and the overviewStore all go away). The
`insightsEntry` preference now mounts/unwinds the registration pair.
Both seats ship on every supported line (0.1.5-rc.1+), pinned per
baseline by the new insightPage seam probe; runtime-verified on a
disposable 0.1.5-rc.1 profile.
The chrome mirrors the shipped first-level pages: one scroll region
with the 960px centered column, the 20px page heading beside the range
group and the balance capsule, and the page element as the `lc-ov`
query container. Inline-size containment makes the page the containing
block for fixed descendants, so the heatmap/usage/balance tooltips can
no longer ride the harness Tooltip (its portal prop is 0.1.7+): they
now carry `data-lc-tip` served by the page's own hover tip — one fixed
bubble portaled to <body>, delegated off the page root (the image
lightbox's precedent), correct on every supported line.
Session card jumps need no close (openSession's own navigation returns
to the conversation); the settings row jump unmounts the page by
selecting the Plugins panel. Verified across 1440-390px in both
locales, and the /context modal is untouched.
The Duration toggle (added in 01b1532) never joined the title-adjacent
zero-margin rule, so its bare .lc-gran kept margin-left: auto and split
the free space with the control cluster — on wide cards it drifted to
the middle of the title row (242px each side at ~930px), and on wrapping
widths it landed wherever the auto margin happened to put it (line 1's
right end in English, line 2's lead at the narrowest).
DNA / Adaptive / Duration now ride the title text's right as ONE
multi-toggle segmented group (.lc-trend-mods), the tips moving from the
group wrappers onto the buttons; the row keeps a single auto-margin push
(the control cluster), so wrapping drops the two semantic groups whole
instead of scattering the modifiers. Also drop the stale "Total/Delta
disables under DNA" note on the disabled-chip rule — no gran button has
been disabled since DNA+Delta shipped.
The Context preferences card gains its eighth row — the trend
duration curve — served across both settings generations: the V3
`settings.register` schema and the V4 Config form (a `.volatile()`
field, so an edit never remounts the entry). The client settings store
parses it fail-open like its neighbours; the trend card's Duration
toggle seeds its mount state from it (default on), and in-chart
toggling stays mount-local and never writes back.
The trend card gains a title-adjacent Duration toggle: every request's
step active time (`activeMs`) drawn as a curve over the bars, read off
a right-hand quartile axis; turn granularity sums the turn's steps, and
a bar without the figure breaks the line instead of faking a point.
The fold stamps it at `step/end` onto the step's committed request
record: the whole step window (the model call plus its tool runs, per
the agent-loop's in-step `executeToolCalls`) minus the step's user
waits — approval decision intervals (`approval/asked` →
`approval/decided`, a crash-dangling pair closing at the step's own
end) and `ask_user_question` answer windows (the tool's body is the
wait itself, in both the blocking and the timed schema's foreground
phase). Overlapping waits price their clamped union, never twice.
The wire field is additive-optional and bumps the timeline projection's
stateVersion (25 → 26): requests already folded can never gain the
figure later, so cached rows refold from the durable log on upgrade.
V4 harness releases log a `developer/message` event whenever the effective tool registry changes mid-session (first shipped in 0.1.7-rc.2): the message nests under data.message with role `developer`, source kind `tool-registry`, and `tool-addition`/`tool-removal` content blocks, lands on the session surface, and the harness's own token meter prices it like any non-system message. The timeline fold had no case for it, so every such message vanished from the plugin's accounting — injection totals, per-request composition, the live surface, and compaction archives under-reported, and the Context tab's numbers could not reconcile with the harness meter or a compaction shadow price.
The fold now treats the event like an injected user message: the nested message is read structurally (the pinned 0.1.5-rc.1 dependency's deriveEventMessage predates the family), the node prices through the shared estimator with the harness's empty-content-is-zero rule, and the injection record takes the producer kind as its label. Developer messages never count as human inputs whatever a hostile source claims; malformed payloads leave the state reference-equal.
stateVersion 24 -> 25: earlier cached rows omitted these nodes and later events cannot recover them, so on upgrade every timeline row refolds once from the durable log (the dashboard warm-up rebuilds idle sessions). The wire shape and durable logs are unchanged. The V4 baselines now carry the event family, and the real-source compat matrix asserts surface tokens, request composition, warm restore, and a v24-row rebuild on both supported V4 lines.
Validation: 102 test files / 1,910 tests with per-file coverage 100%; the real-source matrix passed for 0.1.5-rc.1, 0.1.7-rc.2, and 0.2.0-rc.2; GitHub PR checks green (486069f fixed the workflow's pnpm setup that had failed every external PR at its first step).
The setup step claimed to read the pnpm version from `packageManager` in
package.json, but this plugin package defines no such field, so
pnpm/action-setup aborted with "No pnpm version is specified" eight seconds
into every external PR's run (#102, #103, #107) — no PR check ever reached
the test step. Pin the version explicitly, mirroring release.yml.
Cordis hands out a fresh traced proxy on every ctx.tools property read of a
tracked Service (ToolRuntime is one), each closing over the reader's context.
The hook keyed its WeakSet dedupe on the proxy identity, so every read
re-wrapped and pinned one restore closure per read — patched[] then retained
each caller's whole agent (scope tag -> agent -> session log) past disposal,
and the unload restore guard (instance.register === wrappedRegister) never
held through a traced proxy, leaving the dead wrapper on the shared service.
Peel to the stable underlying instance before dedupe/patch/restore (the
handler still returns the caller's own proxy, keeping scoped registration
this-binding), record the reader as a scalar fiber name instead of the
reader Context, and add a tracked-service regression test: single wrap
across reads, live scoped attribution, and restore on unload. #108
A redacted reasoning block leaves no chunk behind, so the first token
lands after the reasoning window — pricing ttftMs/genMs off the token
charged the marker-tiled decode window to the wait as well, and the
legend's rows summed past 100%. The wait/generation boundary now sits
at the first observable instant (the earlier of the first token and the
first block marker), the same boundary the strip's spans already paint;
the TPS chip keeps the first-token anchor for harness parity.
stateVersion 23 → 24 refolds cached rows with the corrected totals.
The v0.62.0 publish wedged in npm's staged pipeline (the registry accepted
it, signed provenance, then never completed processing — publicly invisible,
and a re-publish is refused with E409 'previously staged'). Same content,
new version to route around the stuck stage.
Folding 1_001 sparse steps (3_003 events, each flushing into the growing
spans array) passed locally but timed out on the release runner. Drive 41
denser steps instead — 50 spans each via a 48-marker decode run — so the
same 2_050 → 2_000 tail-cut assertions hold on 25x less fold work.
The timing card gains a bottom band that reads where the session's ACTIVE
time went: every completed step's slices — the TTFT wait, the decode blocks
in stream order, the tool-run windows, and the in-step residue — packed
gapless in occurrence order, each band's width its share of the cumulative
active time (idle time between steps takes no track), with ticks at zero,
the quartiles, and the full span.
Host: the fold tiles each step at step/end (clamped into the step window,
first-wins de-overlap so parallel runs paint their union, holes filled with
the residue kind) into a new spans collection — bounded to the newest 2,000,
riding the detail channel and the inline view, never the slim head. The
decode window anchors at the first OBSERVABLE instant (the first token, or
an earlier block marker the token packing could not stamp), so a redacted
reasoning block keeps its span instead of being swallowed by the wait.
stateVersion 22 → 23: the per-step collection cannot be backfilled by later
events, so cached rows refold from the durable log.
Client: the strip joins the card's hover link by KIND — a band hover lights
every span of its kind plus the donut arc and the legend row, and the
floating tip answers only the strip's own pointer (the mirror rule).
In DNA+delta combo mode the tooltip's no-band fallback read
tip.total over the delta record, whose total IS the churn —
the label claimed a total while the figure was the change
magnitude (and a zero-churn bar showed 'Total ≈ 0' anywhere).
The band lookup now sits ahead of the mode branches: a hover
with no band under the pointer falls through to the mode's own
row — the shared delta row (fmtSigned of the net) in delta
modes, the total row otherwise. The delta branch reuses
fmtSigned instead of its inline sign formatting.
The DNA on/off state lifts no further than it already sat: the Context
tab's `dna` state now feeds the Context browser through a controlled
pair of props (dna + onDnaChange), so toggling either card's switch
moves the other — the trend bars and the browser's composition bar
become fingerprints together, and drop back together.
The browser stays self-contained where no parent links it: the pair is
controlled only when BOTH props arrive, a lone `dna` prop is ignored,
and the /context modal keeps its mount-local toggle (the established
optional-seam idiom).
DNA and Delta are orthogonal now: the Total/Delta switch stays live under
DNA, and delta diffs each bar's bands against the previous bar's by key
(deltaBandsOf, dna.ts) instead of diffing category totals. Newcomers and
grown items ride the up arm in read order, removed and shrunk items hang
on the down arm in the previous bar's read order, and unchanged items
vanish from both — the first bar carries no change at all, mirroring the
record delta so the axis stays change-driven.
DnaDeltaBar renders each bar as ONE gradient div spanning both arms
around the zero line (down arm reversed bottom-up, then the up arm),
absolutely positioned like the segmented delta arms, with the entrance
rise anchored at the zero line via a per-bar transform-origin. The
cross-bar lifetime highlight becomes a dumb resolved-slice layer both
DNA flavors feed; the tooltip shows the signed per-item change; band
clicks reveal the item in the Context browser through the same bridge
in both flavors. Adaptive rescaling measures the band arms and
re-measures on any mode flip so the axis never rides a stale scale.
A click inside a DNA bar hit-tests the same pointer-to-band mapping the
hover uses and lifts (bar seq, band) to the view, which forwards it
through the browser's one-shot reveal bridge: the browser switches off
the live surface to the bar's step, opens the item's category and
element row, and scrolls it into view. The bar's own click still pins
the step in the detail panel, and both land on the same seq.
The nodeFocus bridge generalizes from message-only (seq) to the band
keys both DNA surfaces share — 'sys' (system prompt), 'tool:<name>'
(a tool schema), 'n<seq>' (a message) — so header bands and message
nodes reveal through one path, and the step-brief / file-activity
locates convert to the same keys. The reveal now also clears the row
lens, so a stale filter can no longer hide the revealed row (the
brief flow had the same latent trap).
DnaBar's hit-test becomes a total function: the first visible band
starts at off 0 (leading zero-token bands advance nothing) and the
pointer fraction clamps into [0, 1], so a hit always exists.
The Context Dashboard entry now renders the document sheet in the
surrounding text colour instead of the six-colour palette, so it sits
quietly beside the harness's own footer rows. The identity seats (tab
chip, command, sidebar registration, panel head) keep the polychrome
default; the package icon.svg is untouched.
Closes#95
Each bar redraws as that request's context fingerprint — ONE gradient
div paints the assembled items in read order (system prompt, tool
schemas, then messages by seq), coalescing adjacent same-color bands,
so thousands of bars cost the same DOM as the segmented stacks.
Hovering a band lights its lifetime slice in every bar at once (the
items' keys are stable across requests), and the tooltip names the
item via the label core shared with the browser's DNA mode
(dnaBaseLabel). The toggle rides the title's right, left of the
adaptive switch; the Total/Delta switch suspends while DNA is on —
DNA plots absolute composition, and a stale 'delta' state degrades to
total semantics instead of emptying the chart. The browser category
focus is ignored (the bands already are the full composition).
Bands are built per displayed record with the same pure client-side
assemble() reconstruction the Context browser uses, memoized on the
timeline so hover re-renders never reassemble. Host-side is untouched
— no wire or protocol change.
Add dsh-v0.2.0-rc.2 to the supported set alongside the 0.1.5-rc.1 and
0.1.7-rc.2 lines: the manifest's dshReleases, the baseline matrix, the
release workflow's tag fetch, and the compatibility docs.
Every seam the plugin rides is spelled as at 0.1.7-rc.2 - the projection
registry contract, the platform module table, the right Sidebar seats and
guide-entry contract, the session-jump verb pair, the detail channel RPC
faces, the Plugins page card slot over the configForms transport, the
balance settings faces, the step-boundary guard, and the durable event
vocabulary (still the V4 log generation, no v4-to-v5 migration). Cordis
stays 4.0.4 and the dsh peers stay >=0.1.5-rc.1.
Verified on the new tag: all 25 compat-matrix probes pass against the real
harness sources and the vendored cordis, `pnpm run lint:fix && pnpm run
test && pnpm run build` is green (1837 tests, 100% per-file coverage), and
on a disposable DSH_HOME the 0.2.0-rc.2 CLI installs dsh-context@0.60.0
into exactly one composed config row and removes it back to zero.
Also refresh the docs' verification stamp (plugin 0.58.0 -> 0.60.0).
A 'Context Browser · Diff baseline' row leads the tool-sort preference:
'prev step' (default) or 'prev turn', reusing the toolbar's vocabulary.
The browser's baseline toggle mounts from the persisted value; in-toolbar
toggling stays mount-local. Served through the settings namespace schema
and the Config-form generation, degrading to the default on stale values.
Replace the fixed 'vs previous turn' hint with a baseline toggle (default:
previous step). A predecessor-less step or live surface diffs against zero,
so the whole makeup reads as change; on live, 'previous turn' measures this
turn's growth so far. The step picker options drop their time suffix.
Add useTimelineSource cases for the cold read rendering from its own slim
head, the retry after a failed/absent/headless cold read, and a pushed head
taking the shared per-session store back. Surface the retryable failure note
in the Context tab and /context modal specs, isolate the page-lifetime
detail stores between tests, and record the no-stall contract in
docs/compatibility.md.
A session can sit with no pushed contextTimeline value at all — the enable
recomposition raced its open baseline (the projection store clears a key the
baseline omits), or its checkpoint row is missing or version-stale on a
cold-observed session. The old code returned a loading surface in that case
with no request and no failure path, so the Context tab and /context modal
stalled on "Reading the session log…" until a refresh.
The source now opens the detail channel itself (rev 0): the detail route
folds the durable log on demand and serves the slim head beside the
collections, the next pushed head takes the channel back per its own
revision, and a read that settles without data surfaces the retryable
failure note instead of an eternal spinner.
Align the assistant text section head with the Answer filter capsule
term, matching dsh's finalized-answer vocabulary for the non-thinking
reply content.
A tool-name capsule — a row tag (tool results, call breadcrumbs) or a
call head (→ bash, ← bash) — no longer clicks through to the tool's
definition row in the Tool Schemas category: the schema reveal from
5a62b2b is removed whole (the reveal jump, the onSchema threading, the
Section label click affordance, the hover-link styles, and the tooltip
key in both locales). The ×N call-name grouping stays — one capsule per
distinct call name, repeats folded into a multiplier — as plain tags on
the row button, so a tag click bubbles to the row's own toggle.
The assistant search box scanned the tag, preview, and the node's
stamped text — the reasoning blocks live only on the conversation join,
so thinking text was unsearchable. Derive one concatenated reasoning
face per assistant row (both block vocabularies, malformed blocks drop)
and scan it alongside the other faces; the placeholder names the new
face in both locales.
A call breadcrumb no longer renders one capsule per call joined by '›'.
The names group by identity: one capsule per DISTINCT tool in
first-appearance order, repeats folded into a multiplier ('bash ×3'),
each capsule still clickable to reveal the tool's schema. The row's tag
slot becomes a shrinkable capsule group, and the hover affordance moves
from the inner segment to the whole capsule. The tag's scan text keeps
the raw joined names, so the text filter still matches every call.
A tool-name capsule — a row tag (tool results, call breadcrumbs) or a
call head (→ bash, ← bash) — is now clickable and reveals the tool's
definition row in the Tool Schemas category: the category opens
unfiltered, the named row expands, and a one-shot scroll lands on it
(the same reveal the DNA bands use). A breadcrumb renders one clickable
segment per call; its segment click stops propagation so the row toggle
cannot overwrite the reveal. Tags that do not name tools stay inert.
Hover affordances: tag segments lift to the primary label on the shared
interactive hover face; call heads shift to the brand face + underline.
The tag capsules (browser row tags, file-activity op tools, step-brief
tags/facts/chip tags) bordered on border-l1 — 6% white in dark mode,
near-invisible against the row. Step the edge to the harness's next
border tier (border-l2, 12% white / 10% black), the pairing the pick
control already uses, so the capsules read as pills in both locales and
themes.
The fold stamps SurfaceNode.calls only on text-less replies, so a reply
carrying both text and tool calls passed the Tools kind filter (proved
via the conversation join) but showed no tool name on its collapsed row.
Fall back to the join-derived call names for the row tag — the same
recovery the step brief applies — so every row the Tools filter keeps
carries its tool breadcrumb.
Add @stylistic/object-property-newline to the src stylistic block, so an
object that already spans multiple lines cannot pack several properties
onto one line. Objects written entirely on one line stay compact, which
keeps small inline props objects untouched.
The autofix split 199 packed properties across i18n.ts, statsTiming.tsx,
fileActivity.ts, index.ts, fold.ts and imageTokens.ts. Note that this
rule's fixer only inserts the newlines, so @stylistic/indent repairs the
resulting indentation on a second lint:fix pass.
The assistant category's search box gains Thinking / Tools / Answer chips, each carrying the shown step's message count; clicking one keeps only the replies carrying that kind (intersected with the text query), clicking again clears, and the pick follows the text filter's lens rules (reset on a category switch, kept across step picks). Kinds classify from the conversation join's blocks (reasoning / tool-call / text, legacy content joins included) and fall back to the surface node's calls/text when the join misses.
A seeded fork (a dsh-better-sidebar Side tab, a subagent_fork child,
a UI fork) replays its parent's event log verbatim, usage settlements
included, so its cumulative cost fold re-priced history the parent
session had already booked - a family of idle Side tabs billed the
parent's whole history again (#94).
The fold now zeroes the session-cost totals at the tagged
`session/end-seed` marker the session core stamps at the inherited
cut, in both projection units: the timeline's cost book and the
activity day ledger (whose armed step slot dies with the cut, so a
child settlement cannot attribute to a parent day). Untagged resume
markers reset nothing; nested seeds re-mark at their own cut, so a
session's totals are always its own post-seed spend. stateVersion
bumps (timeline 22, activity 3) refold cached rows from the log on
upgrade.
bowenliang123