---
title: "Claudoscope changelog: every release of the Claude Code menu bar app"
description: "Every Claudoscope release, newest first. Version 1.3.0 shipped October 4, 2026."
canonical: https://claudoscope.com/changelog
last_updated: 2026-10-04
---

# What changed in Claudoscope

Every release, newest first. The current version is 1.3.0, released October 4, 2026. DMGs for each version are on [GitHub Releases](https://github.com/cordwainersmith/Claudoscope/releases).

## Claudoscope 1.3.0

*October 4, 2026*

Adds a Fleet view: one board for every Claude Code session that is running now or ran in the last 24 hours, across all projects, with an attention queue for the ones waiting on you. Built on the session registry Claude Code writes at `~/.claude/sessions/`, which Claudoscope had never read, so liveness comes from the process itself rather than from how recently a transcript grew. Also catches Claudoscope up with Claude Code 2.1.271 through 2.1.289, including a billing correction for Opus 5.5.

Requires a one-time full reparse on first launch (parser version 13), after which launches hydrate from cache as before.

### New Features

- **Fleet rail.** A full-width board instead of the usual list and detail columns. A headline says how many agents need you, with live, working, waiting and skipped-permission counters beside it. Below that, a "Needs you" lane lists waiting agents oldest first with a ticking wait timer, the reason, and Open and Jump buttons (Command+Return jumps to the oldest). Working and Parked lanes show live agents as cards, and a dimmed Landed list covers everything else from the last 24 hours. Clicking any card slides in an inspector with the agent and process details, Focus terminal and Open session. State precedence is hook event, then Claude Code's own registry status, then transcript recency, so a card's state always traces to a concrete record. A filter field and a skipped-permissions toggle narrow the board.
- **Live stats on every card.** Each card shows time in its current state and the session's age, your last prompt as its title, the last tool call and what it touched, a context gauge for the latest turn (amber and red by share of the window and by absolute size, so a 1M window cannot hide a bloated context), average spend per hour, cache hit rate, and chips for compactions and errors. Idle and waiting agents show how long their prompt cache has left, so you can answer the one that is about to go cold before it costs a full re-cache.
- **Fleet telemetry.** A strip under the headline reports today's spend by the sessions on the board, the combined spend rate of live agents, and an hourly activity chart for the last 24 hours.
- **Menu bar waiting count and fleet state in the popover.** The menu bar icon shows how many agents are waiting on you, and turns its dot red when a live agent ran with skipped permissions. The popover's Active Sessions card now carries each session's fleet state: waiting and blocked rows come first with a live wait clock and the reason, the card tints orange or red when something needs you, and its header opens the Fleet view.
- **Jump shortcut.** An optional system-wide Control+Option+Command+J brings forward the terminal tab of the agent that has waited longest, or opens the Fleet view when nothing is waiting. Off by default; enable it under Settings, Fleet. No Accessibility permission needed.
- **Skipped-permissions marking.** A card carries a red Bypass tag while its session is running with `--dangerously-skip-permissions`, and a faint shield if it bypassed earlier and has since left the mode. Derived from the transcript's permission-mode records. Cards also show the current permission mode and git branch.
- **Live sessions in the popover.** The active-sessions card now counts a session as active while its process is alive, not only while its transcript is changing.
- **Notification taps select the session.** Clicking a "Claude needs you" or "Claude is ready" notification still focuses the terminal, and additionally selects that session in the dashboard when the window is open.
- **Popover rows act on click.** Clicking a waiting row in the popover's Active Sessions card brings its terminal tab forward; clicking any other row opens that session in the dashboard (Cowork rows open the Fleet view).
- **Crash detection.** When a process disappears from Claude Code's registry while it was reported busy, and no Stop hook arrived, its card lands as Failed with "Process exited mid-turn". Resuming the session and sending a prompt clears it.
- **Long waits escalate.** A wait longer than ten minutes, or one whose prompt cache is about to expire, turns the menu bar count, the popover row and card, and the board tile red. This is visual only; no extra notification is posted.
- **Response time and cold-cache cost.** The telemetry strip now shows how long agents waited on you today (mean and longest) and how many answers came after the prompt cache expired, with an estimate of what re-caching the context cost. Wait history is kept locally for 30 days.
- **Copy resume command.** Landed rows (on hover) and the inspector copy a ready-to-paste `cd <checkout> && claude --resume <id>` line for sessions whose process has exited. Claudoscope never runs it.
- **Session names.** Cards, landed rows, the inspector and popover rows show the session name Claude Code gives a running session (its auto topic name or a rename), and the filter matches it.
- **Refused-call and changed-file chips.** Cards show how many tool calls you or a permission rule refused and how many files the session edited. Clicking the refused chip opens Chat with the blocked actions expanded; clicking the files chip opens the Files tab.
- **Group by project.** A toggle in the Fleet header groups cards and landed rows under project headings. The choice is remembered.
- **Per-agent budgets.** Set a dollar budget for a session in the Fleet inspector. When cost alerts are on, crossing it (and each doubling after) posts a notification whose tap selects the session and focuses its terminal, even if the general per-session rule is off. A budget chip on the card turns amber at 80% and red at 100%.
- **Fleet tools for the MCP server.** Two new read-only tools, `list_agents` and `get_agent`, return the Fleet board's agents, states, wait reasons and times, process details, cost and context, with agents that need you first.
- **Cost by plugin.** Claude Code now tags billed turns with the plugin that drove them. Analytics > Attribution gains a Cost by Plugin table, and each plugin's detail shows its lifetime spend. Like skill and MCP attribution, it is a partial partition with its own unattributed remainder.
- **Claude Mods in the Plugins rail.** Plugins that ship TypeScript mods (2.1.287) carry a mod badge, and their detail lists the events each mod registers, flagging the ones that can change behavior (tool calls, prompt submission and composition, session appends). Found by reading the source; nothing is executed.
- **AGENTS.md as project memory.** A project with no CLAUDE.md has its root AGENTS.md shown in the Memory rail and linted.
- **More of settings.json surfaced.** `"attribution": false` (2.1.281) is shown as all attribution hidden, and `deniedModels` and `availableModelsMatch` are displayed, managed values winning over user ones.

### Fixed

- **Opus 5.5 priced at its own rate.** `claude-opus-5-5` was billed as Opus 5 ($5/$25, cache reads $0.50), overstating its cost by a quarter. It now bills at $4/$20 with $0.20 cache reads, with cache writes and the Vertex rows derived from those. A managed pricing override for one of Opus 5 or Opus 5.5 no longer reprices the other.
- **Plugin hooks are visible.** Plugin `hooks/hooks.json` files that wrap their events in a `hooks` key, which is how real plugins ship them, were read as empty, so those hooks were missing from the Hooks rail, HOOK lint, and the MCP server's `get_config`. Hooks defined as a command plus arguments now show as the full command line and are matched in the Hooks Runtime tab.
- **More refused tool calls are counted.** Only three denial phrasings were recognized, so auto-denials, safety-check blocks, output-redirection blocks and bare rejections were missing from blocked-action counts and Chat's blocked-actions section. All current forms are now recognized.
- **Fleet cost includes subagents.** The board's spend, burn rate, card costs, budgets and the fleet MCP tools counted only the main transcript, so a session that delegated heavily looked far cheaper than the menu bar said. Subagent spend, nested ones included, is now folded into each agent, and the inspector shows how much came from subagents.
- **Quiet agents no longer stick at Working.** A session with no live process (Cowork, or a CLI process whose registry file is gone) stayed in the popover card and the Working lane after its transcript went quiet, because the board was only re-derived while a CLI process was live. The board now keeps re-deriving until such agents age out.
- **Refused tool calls are parsed again.** Transcript lines where Claude Code records a refused or invalid tool call as a plain string were dropped whole by the parser, so Chat's blocked-actions section never showed real refusals. Subagent results with block content were dropped the same way, which lost the link from a session to the subagents it spawned. Both now parse.
- **A second copy of Claudoscope can no longer leave stale data in the cache.** If an older build was still running when a newer one started, the older one could write session summaries in its own format into the newer build's cache, and the newer build trusted them on its next launch. Each cached summary now records the parser version that wrote it, and anything from another version is discarded and reparsed.

### Changed

- **Hook events feed fleet state even when notification delivery is off.** The Notification and Stop hook spool is read for board state before the notification service drains it. The hooks are still installed only when Notifications are enabled; the Fleet view says so and links to that setting.

### Security

- **New CFG022.** A boolean `attribution` in user or project settings.json. Claude Code versions before 2.1.281 skip a settings file containing it, so everything else in that file silently stops applying.
- **New CFG023.** OpenTelemetry export, endpoint or content-capture variables set in project or project-local settings, which Claude Code ignores since 2.1.282.
- **New CFG024.** An MCP server named `anthropic-skills`, a name reserved since 2.1.282.
- **New CFG025.** An MCP entry of type `sdk`, which Claude Code skips (2.1.274).
- **New CFG026.** `deniedModels`, `availableModelsMatch` or `allowedProviders` in user settings, where they are documented as managed-only.
- **New HOOK005.** An agent-type hook on `PermissionRequest` (2.1.280).

## Claudoscope 1.2.0

*September 17, 2026*

Catches Claudoscope up with Claude Code 2.1.238 through 2.1.270. Claude Code now stamps transcripts with which skill, MCP tool, or subagent drove each billed turn, which turns three configuration rails into spend surfaces: the Skills, MCPs, and Agents rails can finally answer what a thing costs, not just whether it is installed. Alongside that, two cost corrections you could not have spotted from inside the app: organizations on contracted rates were shown list price, and the pricing provider silently reset to Anthropic on every launch.

Requires a one-time full reparse on first launch (parser version bump plus a pricing-table key change), after which launches hydrate from cache as before.

### Fixed

- **Contracted rates are honored instead of list price.** Claude Code 2.1.243 lets an organization publish negotiated per-model rates through managed `modelPricing`. Claudoscope read none of it, so every cost figure on a machine under a contract was the public price. Overrides are now resolved once, before the first scan, and flow through the single pricing table the whole app already shares, so every cost, chart, and alert reflects them. A row written by an override suppresses the fast-mode 2x multiplier, since the documented behavior is that the written rate is used as-is, and the resolved table is hashed into the cache key so adding, editing, or removing an override rebuilds costs rather than leaving stale ones in place.
- **The pricing provider no longer resets to Anthropic on every launch.** `pricingProvider` and `pricingRegion` were never persisted. Picking Vertex held for the session, then reverted at next launch, and re-picking it wiped and reindexed the entire summary cache each time. Both now persist, so the choice survives a relaunch and costs no longer silently revert to a different provider's rates overnight.
- **Fable 5.1 cache reads priced at $0.25/MTok.** `claude-fable-5-1` became the default Fable model in 2.1.257 with cache reads at a quarter of Fable 5's $1.00, while input and output stay at $10/$50. It was previously billed at the Fable 5 cache-read rate, overstating cost on exactly the token class a long session accumulates most of.
- **Cowork sessions no longer lose their hook runtime stats.** The Cowork parser had been dropping the field since 1.1.0 introduced it, so hooks that fired inside a Cowork session were missing from the Hooks Runtime tab. It now forwards every pass-through field.
- **Hooks referencing a managed MCP server are no longer reported as broken.** Managed servers were invisible to Claudoscope, so HOOK003 checked hook matchers against a server list that omitted them and flagged working configuration. Managed servers are now part of that inventory.

### New Features

- **Cost attribution across skills, MCP tools, and subagents.** Claude Code 2.1.24x tags billed records with the skill, MCP server and tool, or subagent type that drove the turn. A new Analytics tab reports spend per skill, per MCP tool, and per agent over the selected window, and the Skills, MCPs, and Agents rails each carry a lifetime spend chip. Skill and MCP attribution are independent partial partitions rather than two slices of one pie (a single turn can carry both tags), so each is shown with its own unattributed remainder and neither is normalized to 100%. Coverage is roughly a tenth of billed turns, because only transcripts written by 2.1.24x and later carry tags at all, so every surface states its coverage; a rail entry with no attributed spend shows no chip rather than `$0.00`, since absence cannot distinguish "never ran" from "ran before tagging existed". The same skill is tagged both bare and plugin-qualified, and both forms fold to one row.
- **Prompt cache health.** Cache Effectiveness gains a card reporting tokens re-primed and what that cost, a cold-turn floor, and the two prefix-rebuild causes a transcript can actually evidence: a session resumed on a later day (no cache TTL survives overnight, so each later day rebuilds the whole prefix) and a mid-session model switch (one model cannot read another's cached prefix). The headline is the wasted 1-hour TTL: a session that wrote heavily on the 1h tier and read back less than it wrote paid roughly twice the input price for nothing, and the premium over the 5-minute rate is reported per session. Everything here is derived from per-day and per-model totals rather than measured per turn, so the card is labelled inferred and the counts are stated as lower bounds.
- **Plugin scope drift.** The same plugin can be installed at both user and project scope at different commits, and which copy a session loads depends on the directory it started in, so two projects can run different code under one plugin name with nothing in the UI distinguishing them. The Plugins rail now reads the per-scope install records, badges a drifted plugin, lists each scope with its version and commit, and explains what the drift means for your specific projects.
- **Managed MCP servers in the MCPs rail.** `managedMcpServers` from organization policy was the one MCP source Claudoscope never read. Managed servers now appear with a `managed` badge and take precedence on a name collision, matching Claude Code, which does not let a user config override them.
- **Subagent model and effort.** The agent tree shows the effort level each subagent actually ran at alongside its model, and a background job's detail card reports the model, effort, and estimated cost of the run rather than only its token count.
- **More of settings.json surfaced.** `maxEffortLevel`, `bashOutputMaxChars`, `taskOutputMaxChars`, `promptCacheTtl`, and `subagentPromptCacheTtl` are read and displayed, and the Permissions section states whether reads are confined to the working directories. The cache TTL keys are display only: cost already prices the 5-minute and 1-hour tiers from each record's own split, so nothing about billing depends on them.

### Security

- **New CFG019.** A `managedMcpServers` entry that defines a stdio command is reported as an error. Claude Code accepts only http and sse servers in managed scope and skips anything else without saying so, so an administrator can believe a server is deployed fleet-wide while it runs nowhere.
- **New CFG020.** A `permissions.allow` Bash rule whose wildcard is not its final segment is a privilege escalation, not a typo: Bash rules match a prefix, so `Bash(git * main)` ignores everything after the `*` and permits any command beginning `git`, chained ones included. Checked across user, project, and project-local settings, because allow lists usually live in the repo.
- **New CFG021.** `bashOutputMaxChars` or `taskOutputMaxChars` outside the range Claude Code accepts, which it clamps without reporting, so the configured number is not the one in effect.
- **New HRD014.** `permissions.blockReadsOutsideWorkingDirectories` not enabled, which leaves Read, Grep, and Glob able to reach any path on disk, including other repositories and credential files.
- **Containment Escape added to the hardening auto-mode baseline** (2.1.257), covering cloud metadata-credential fetches, egress-allowlist evasion, and cross-tenant reach. Installing the baseline now also satisfies HRD012, which previously reported a missing `hard_deny` block that the baseline never supplied.
- **Four new skill checks.** SKL010 flags a relative link in `SKILL.md` that resolves to nothing, so instructions Claude was meant to load are silently absent. SKL011 flags a tool restriction naming an MCP server that is not configured, where an allow entry grants no access and a deny entry blocks nothing. SKL015 flags the same skill name defined in two scopes, which Claude Code resolves to one without reporting which, so an edit can land in the copy that is not running. SKL016 flags an installed skill that never appears in an attributed turn while others do, since its description occupies context on every request; it stays silent until the corpus holds enough tagged turns for the absence to mean anything.
- **New PLG004** for the plugin scope drift described above.
- **`PreModelSwitch` and `PostModelSwitch` recognized as hook events** (2.1.251), so a matcher set on either is correctly reported as dead config.

## Claudoscope 1.1.0

*August 26, 2026*

Catches Claudoscope up with Claude Code 2.1.196 through 2.1.237, and opens three new observability surfaces: session quality insights, hook runtime stats, and a background tasks and jobs monitor. Two of the fixes are billing corrections you would not have noticed from inside the app: one model priced at zero, and another about to be priced 50% high.

### Fixed

- **Sonnet 5 no longer reverts to $3/$15 in September.** Sonnet 5's $2/$10 launch rate was announced as introductory pricing through 2026-08-31, and Anthropic has since made it the standard price and cancelled the increase. The dated cutoff has been removed, so Sonnet 5 keeps billing at $2/$10 indefinitely instead of jumping 50% on 2026-09-01.
- **Claude Mythos 5 is priced.** `claude-mythos-*` matched no model family, so it resolved to the "unknown" rate of zero: sessions using it contributed nothing to any cost figure and were dropped from the model distribution entirely. It now prices at $10/$50 (Fable 5 rates) on all three tables.
- **Haiku 3.5 billed at Haiku 3 rates.** Both ids shared one table row, underbilling Haiku 3.5 by about 3.2x. Haiku 3.5 now uses its own $0.80/$4 rate; the family label is unchanged, so nothing forks in the UI.
- **Unpriced models are visible instead of free.** Analytics shows a notice naming any model id with no known rate. Previously such a model cost $0 and was filtered out of the model distribution, so a session that ran entirely on one looked like it had never happened.
- **Skill tool restrictions no longer flag working config.** The known-tools list had stalled at twelve entries, so any skill whose `allowed-tools` named Agent, Skill, ToolSearch, AskUserQuestion, SendMessage, or a Task tool drew a spurious "unknown tool name" warning. The list now covers the current tool surface.
- **`DirectoryAdded` recognized as a hook event** (Claude Code 2.1.219), so a matcher set on it is correctly reported as dead config.
- **GitLab marketplaces and archive plugin sources display their source.** `additionalMarketplaces` is now read as an alias for `extraKnownMarketplaces` (2.1.232), and the `archive` (2.1.224) and `command` (2.1.229) plugin sources show their URL or command instead of a blank detail.

### New Features

- **Insights tab in Analytics.** Reads the session facets Claude Code's `/insights` writes to `~/.claude/usage-data/` (outcome, friction, satisfaction, goal, session type) and joins them to Claudoscope's cost engine: outcome distribution, friction frequency, and average cost by outcome, plus a per-session facet detail with a jump to the session. A coverage banner states how many sessions have facets and when they were last generated, since facets exist only after running `/insights`. Lives as a Usage/Insights toggle inside the Analytics rail rather than its own icon.
- **Hooks Runtime tab.** The Hooks rail gains a Configuration/Runtime toggle. Runtime shows what actually executed across all parsed sessions, folded from the hook execution records in transcripts: per-hook fire counts, failures, average and max duration, session counts, and a "not in config" badge for commands seen in old transcripts that match no current hook. Sidebar event rows carry fire/failure counts, and the chat view marks each Stop-hook batch inline, including one that blocked continuation. Requires a one-time full reparse on first launch (parser version bump).
- **Tasks & Jobs rail.** Surfaces Claude Code's background jobs (`~/.claude/jobs/`: state, timeline, tokens, result) and per-session task lists (`~/.claude/tasks/`: checklist with dependency chips), with jump-to-session where the owning transcript is still in the corpus and a daemon status line. Job `providerEnv` maps are never decoded, so credentials in job state cannot surface anywhere in the app.
- **Merged Health rail.** Health, Hardening, and Routing were three icons over the same lint results; they are now one Health rail with a section toggle, shrinking the icon rail.
- **Context tab.** Per-session chart of how full the context window got on each assistant turn, against that model's ceiling, with compaction events marked. Reports peak context and peak utilization, and flags sessions that mix model generations across the Claude 4.7 tokenizer change, where token counts are not comparable turn to turn.
- **Session provenance.** Sessions started with `--worktree` or `/fork` show their worktree and branch, and a session that opened a pull request or GitLab merge request links to it from the session header. Claude Code's own generated session names (`ai-title`) are now used as titles, ranking below a `/rename` and above the slug.

### Security

- **Nine more GitLab token families detected.** Only `glpat-` was recognized; runner, OAuth, pipeline-trigger, agent, import, service-account, CI-build, feature-flag, and deploy tokens now are too. The routable `glpat-`/`gldt-` pair is classified as a critical account-level credential.
- **`glab` credential store protected** in the hardening sandbox baseline, matching the existing `gh` entry.
- **Six new configuration checks (CFG013 through CFG018).** Filesystem isolation disabled (2.1.216); sandbox running without a strict network allowlist (2.1.219); credential `mode: "mask"` set without the TLS termination it depends on, which silently leaves masking inert (2.1.221/.224); sandbox binary overrides in project settings, which Claude Code stopped honoring (2.1.232); `remoteControlAtStartup` in project settings, which can no longer enable Remote Control (2.1.222); and cross-session messages auto-accepted into a session running with bypassed permissions (2.1.224).
- **New SKL014.** Flags a skill restricted only to `TodoWrite` or `Task*` tools, which Claude Code 2.1.233 removed from Opus 4.8, Sonnet 5, Fable 5, Mythos 5, and newer models unless `CLAUDE_CODE_ENABLE_TODO_TOOLS=1` is set.
- **New CHN family for channel plugins (CHN001 through CHN003).** Channels let an external endpoint push messages straight into a live session, and a two-way channel can opt into relaying tool-permission prompts for remote approval, so an installed one is worth seeing on purpose rather than discovering later. CHN001 reports each enabled channel plugin from the research-preview set (Telegram, Discord, iMessage, fakechat) as a prompt-injection and permission-relay surface. CHN002 notes when such a plugin is enabled while `settings.json` selects Vertex or Bedrock, where Claude Code silently ignores channels and the plugin is inert; detection is best-effort, since provider env vars exported only from a shell profile are not visible to the linter. CHN003 surfaces the `channelsEnabled` org-policy key whenever it is set, either way, so its value can be checked against intent. All three land in the Plugins category of Config Health.

## Claudoscope 1.0.0

*July 27, 2026*

🎉 **One point oh.** Claudoscope started as a menu bar readout for a single number and grew into a full lens on where Claude Code actually spends your time and money. This is the release where the data holds still: session summaries persist between launches instead of being re-parsed every time, cost estimates have been audited against real bills, and everything Claude touched in a session is one tab away.

Thanks to everyone who filed an issue, argued with a pricing table, or just left it running in the corner of a screen.

### New Features

- **Persistent session index (SQLite).** Launch no longer re-parses every transcript: parsed session summaries persist in a GRDB-backed cache at `~/Library/Application Support/Claudoscope/cache.sqlite`, keyed by file size and mtime (context-fork subagent files additionally fingerprint their parent transcript). The dashboard paints from the cache immediately, sub-second on a ~3,000-file corpus that previously took ~25 seconds, then a background reconcile re-parses only changed files, with the scan banner counting changed files only. Deleting a session file now removes it from the UI live (new watcher delete event) instead of at next launch. The cache is a pure derivative: a parser change (versioned via `SessionParser.parserVersion`, which MUST be bumped with any billing/parse-logic or `SessionSummary` field change), a pricing rate edit (tables are hashed into the cache key), a provider/region switch, or a timezone change wipes and rebuilds it in the background, and a corrupt or deleted cache file rebuilds silently. Pricing switches and FSEvents overflows reuse the same reconcile path, so neither blocks on a full re-parse anymore. Cost alerts are suppressed until the first reconcile completes and the spend ledger rebaselines, so a week of offline appends never reads as a spend spike.
- **Cost alerts (optional, off by default).** Settings > Cost Alerts adds four global rules with fixed thresholds in estimated dollars or tokens: single-session cap, rolling-window spend (5 minutes to 4 hours, doubling as runaway-burn detection), daily total, and monthly total. Alerts re-fire at each doubling of the threshold (X, 2X, 4X, ...), are delivered as a macOS notification plus a red menu bar dot, and stay in the popover until dismissed. Clicking a notification opens the dashboard. All figures are estimates; fired thresholds persist across relaunches so nothing alerts twice. Rolling spend is tracked by an in-memory delta ledger that rebaselines on rescans and pricing switches, so pricing changes and offline gaps never produce phantom alerts.
- **Session lifecycle notifications (opt-in).** Settings > Notifications installs Claude Code `Notification` and `Stop` hooks and delivers two independently toggleable native macOS notifications (both on once enabled): "Claude needs you" for a real block (permission, plan, or MCP prompt), and "Your turn" when Claude finishes a turn. Both are event-driven, so neither re-fires on its own. Includes per-project mute, daily quiet hours, and a sound toggle. Each notification is labeled with the project/repo folder name (a `/rename` or slug title appears as "Title (folder)"), taken from the hook payload so it reads the same across any terminal. An existing hand-rolled `session-notify.sh` is detected, and its Notification and Stop entries are backed up and disabled on enable, then restored on disable, so banners are never duplicated.
- **Per-session Files tab.** Session detail gains a Files tab listing every file Claude edited or wrote, with chronological per-edit diffs rendered from the `structuredPatch` payloads already in the transcript. Subagent edits are merged in with agent badges and anchored to the call that spawned them. Each edit offers open, reveal, copy-patch, and jump-to-chat, and a badge marks files modified on disk since the session touched them. A dedicated service re-streams the transcript when the tab opens, so the parser, the lite scan, and the summary cache are untouched.
- **Canon rail.** An opt-in, per-project store of settled engineering decisions that travels with the repo instead of living in per-machine memory. Enabling a project installs a protocol rule and a seeded records file into its working tree so Claude Code captures decisions in a file committed alongside the code. The rail displays records read-only as structured cards with a kind filter and a hide-superseded toggle: the app is installer and viewer, never a record writer. Includes bulk enable/disable in Settings and a CAN lint family (missing protocol, gitignored records, malformed record, outdated protocol, dangling supersede) surfaced in Config Health.
- **Agent Routing rail.** Installs a set of cost-aware, role-scoped Claude Code subagents (recon, Explore, routine, builder, checker, plus security-review and security-build) into `~/.claude/agents/`, appends a marker-delimited orchestration policy to `~/.claude/CLAUDE.md`, and sets a `fallbackModel` only when one is absent. Install, reinstall, revert, and uninstall mirror the Hardening rail, with timestamped owner-only backups before every write and a conservative uninstall that preserves agent files you have edited. An RTG lint family flags drift.
- **Agents rail.** A read-only inventory of every agent installed on the system: user-level, per-project, and plugin-provided. Core routing agents are grouped into a pinned section with a badge so it is obvious which ones Claudoscope installed. Building this also fixed a frontmatter parser bug that treated the opening `---` fence as the closing one and silently dropped metadata, which restores frontmatter in the Skills rail and the Plugins drill-down too.
- **Embedded MCP server (off by default).** An optional read-only MCP server runs inside the app so Claude Code can query your own usage data. Nine tools cover usage totals (reusing the same analytics engine as the dashboard, so the numbers match), project and session listing and search, config linting, plans, and canon. It is served over a `0600` unix socket via a bundled stdio shim, secrets are always masked, and enabling it registers the shim through the `claude` CLI rather than editing `~/.claude.json` directly.
- **Focus filter in the chat view.** A Focus toggle with separate Thinking and Tool/MCP switches hides the machinery so you can read just the conversation. Filtering is render-time only, so tokens, cost, and analytics stay computed on the full transcript. Assistant records that render nothing (empty streaming fragments, blank tool-result rows) are now dropped everywhere, and in-conversation search matches only what is visible.
- **Monochrome menu bar icon.** Settings > Appearance can switch the menu bar icon to a template-rendered glyph that adapts to light, dark, and tinted menu bars. The color icon's alpha channel is too soft for template rendering, so this uses a separate crisp silhouette asset rather than a desaturated copy.

### Improvements

- **Persistent project and date filter.** A global lens above the sidebar scopes the sessions, tools, timeline, and plans rails at once and survives switching between them, unlike the per-rail text filter. It reuses the Analytics date-window semantics but keeps its own state, so the Analytics controls are unaffected.
- **Session notifications focus the terminal.** Tapping a "Claude needs you" or "your turn" banner now focuses the terminal tab running that session, restoring the click behavior of the hand-rolled script it replaced. Supports Ghostty, iTerm2, and Terminal.app by title-matching the project folder. This requires the Apple Events entitlement, so the app asks for automation permission the first time you tap a banner.
- **Settings button in the menu bar popover**, so the popover no longer has to route through the dashboard window to reach preferences.
- **Redesigned About screen** as a proper standalone window.
- **Recent sessions no longer mirror Active.** The popover's Recent list was sorted purely by recency and took the top three, so active sessions (the most recent by definition) always duplicated the Active Sessions card above them. Active sessions are now excluded before the cut, so Recent complements Active instead of repeating it.

### Bug Fixes

- **Opus 5 priced at 3x the real rate.** `claude-opus-5` is the first opus id with a one-part version, and the version parser only matched two-part ids such as `opus-4-8`. Detection fell through to the pre-4.5 `opus4` family, so every Opus 5 session was billed at $15/$75 per MTok instead of $5/$25, inflating session, project, and analytics totals threefold (one 82M-cache-read session read as $151.80 instead of $50.60). Model-family detection has been inverted to fix the underlying fragility, not just this instance: instead of parsing a version and treating anything unparseable as legacy, it now matches an explicit closed list of the generations that actually billed at the older rate (Claude 3 Opus, Opus 4 and 4.1, Haiku 3 and 3.5). An unrecognized model id therefore prices at the current rate rather than silently inheriting a 3x legacy one. Requires one full reparse, which happens in the background on first launch.
- **Web-search request fee was never billed.** Claudoscope under-counted cost against `/usage` on turns that used web search. The fee is $0.01 per search, but Claude Code records the count in `toolUseResult.searchCount` on the tool-result record rather than in `usage.server_tool_use.web_search_requests`, which is always zero in transcripts. The fee is now billed per search, deduped by record, and attributed to the day and model that issued it, with the popover and Cowork totals kept at parity.
- **Sidebar froze at 100% CPU on large corpora.** The global filter checked each project for surviving sessions by calling a getter that recomputed the entire projects-by-sessions map from scratch every time it was accessed. At real scale (74 projects, ~3,000 sessions) a single sidebar render performed roughly 230,000 timestamp parses and hung the UI. Both accessors now share one per-project helper, taking that render from an 80-second freeze to 0.3 seconds, with a regression test sized to the same corpus.
- **Canon enable no longer pollutes nested projects.** "Enable Canon for all projects" wrote `.claude/rules/canon.md` into every tracked project directory, including ones that are not real project roots: a parent folder that was once a session cwd (e.g. `~/projects`), or a subdirectory of a repo. Because Claude Code loads `.claude/rules/*.md` hierarchically, a rule at an ancestor directory reappears inside every descendant project as a duplicate. Canon now installs only at eligible targets: git repository roots and standalone non-git folders. A directory that contains other tracked projects (a container) is skipped, and a repo subdirectory folds into its repository root, so canon lands once per repo and never cascades. Bulk enable reports how many projects were skipped as ineligible, per-project enable refuses containers and subdirs with an explanatory message, and the Canon rail lists only eligible projects. Opt-in state left by the old behavior self-heals: a project that is now classified as a container or subdir has its stale opt-in dropped on the next Canon refresh.

## Claudoscope 0.9.0

*July 6, 2026*

### New Features

- **File-history checkpoint timeline.** The chat view activates the previously inert `file-history-snapshot` records: a collapsible "Files changed (N)" section lists each file touched in the session with its latest version and last backup time, and assistant turns that produced a backup gain a "Checkpoint" marker. Reads the in-transcript snapshot index only; reading `~/.claude/file-history/` for diffs is a later phase.
- **Blocked and denied actions in chat.** Auto mode (Claude Code 2.1.183+) blocks destructive git and IaC commands and routes denial reasons into the transcript, but there is no dedicated on-disk denial record. Claudoscope now detects the canonical `is_error` tool-result markers and classifies them (destructive git, IaC destroy, permission-setting, user-rejected) into a collapsible "Blocked & denied actions" section. Full-view only, since embedded tool-result denials are invisible to the lite scan.
- **MCP auth status in the MCPs rail.** Claude Code 2.1.186 added OAuth for remote MCP servers. Each server now shows a best-effort auth state derived from its transport plus the `~/.claude/mcp-needs-auth-cache.json` hint (timestamps and hashes only, no secrets): stdio is n/a, an http server flagged in the cache needs login, an unflagged http server is treated as authenticated. The sidebar status dot is colored accordingly and http server cards show an auth capsule. Deliberately file-only to preserve the app's no-secrets posture, since OAuth tokens live in the Keychain.
- **Launch at login.** An optional login item registered via `SMAppService`, gated behind a one-time prompt after onboarding and a reversible toggle in Settings > General. Off by default; the system login-item status is the source of truth.
- **Hook-matcher linter (HOOK001 through HOOK004).** A new HOOK rule family flags hook matchers broken or weakened by recent Claude Code changes: an `mcp__` matcher with no tool segment that now matches nothing (HOOK001), a comma-separated matcher that silently never fired before 2.1.191 (HOOK002), a matcher targeting an MCP server absent from the loaded config (HOOK003), and a matcher set on an event that ignores matchers (HOOK004).
- **New settings.json keys surfaced and linted.** Recognizes the June 2026 Claude Code keys (`sandbox.credentials`, `sandbox.allowAppleEvents`, `attribution.sessionUrl`, `autoMode.classifyAllShell`, `availableModels`, `enforceAvailableModels`, `requiredMinVersion`/`requiredMaxVersion`, `respondToBashCommands`) and adds four governance lint rules: CFG009 (sandbox enabled without credential denies), CFG010 (Apple Events weaken isolation), CFG011 (`respondToBashCommands` off while hooks expect it), and HRD013 (models listed but not enforced).

### Improvements

- **Cache Coverage stat.** The Cache analytics view adds a Cache Coverage figure next to Hit Rate: the share of all input tokens served from cache (reads ÷ reads + writes + fresh input). It sits below Hit Rate and drops sharply when a turn sends large uncached content, whereas Hit Rate (reads ÷ reads + writes) can stay high. The dollar impact remains on the Savings card.
- **Depth-aware subagent trees.** Claude Code 2.1.172 lets subagents spawn subagents up to five deep, but the tree builder only ever produced one flat level. It now reconstructs the full forest recursively from the `toolUseResult.agentId` edges (with a cycle guard and depth cap), falling back to flat rendering when no edge info is present.
- **Colorblind-safe chart palette.** Analytics charts previously used raw system colors, putting green, orange, and red in the same chart and collapsing for red-green colorblind viewers. Charts now share an Okabe-Ito categorical palette, a cool-to-warm ordered scale for ranked data, and a single `Color.forModel` helper, with separators added between donut slices.
- **Sonnet 5 pricing.** `claude-sonnet-5` resolves to the existing Sonnet family at $3 input / $15 output (Sonnet's rate has been flat from 3.5 through 5), so no new table row is required; regression tests now lock the family detection, Anthropic and Vertex-regional rates, and an end-to-end parse.

### Bug Fixes

- **Chat CPU spin on large sessions.** Opening a large, live-growing session pegged the main thread at 100% CPU because `parseMarkdown` ran from a computed property on every SwiftUI body evaluation for every visible and prefetched message. Parsing is now memoized (an `NSLock`-guarded LRU, capacity 512) and each view parses its blocks once in an explicit init.
- **Per-session chat state bleed.** `ChatView` had no view identity, so switching sessions reused the same instance and its state (search text, highlights, scroll flags) leaked from the previous session. Adding `.id(session.id)` resets state on session switch without tearing down on file growth.

## Claudoscope 0.8.0

*June 11, 2026*

### New Features

- **Plugins rail.** New rail that inventories installed Claude Code plugins and the components each one contributes (commands, skills, agents, hooks, MCP servers). Expand a plugin to see component names grouped by kind, and click any component to open its source (a skill's `SKILL.md`, a command or agent `.md`, or a hook/MCP config JSON) in a sheet. Three dependency lint rules back the rail: PLG001 (a plugin declares a dependency that is not installed or enabled), PLG002 (plugin dependencies form a cycle), and PLG003 (a plugin contributes no components).
- **Fable 5 model family.** `claude-fable-5` is now recognized and priced; previously it fell through to the "unknown" family, billing $0 with an "Unknown" badge. Added to all three pricing tables ($10 input / $50 output; Vertex regional $11/$55, marked provisional), rendered in pink across the charts, agent tree, and cache view, listed in the settings pricing table, and flagged as expensive in the efficiency table.
- **Cowork in the menu bar.** The popover's Today stats (Sessions, Tokens, Cost), Active Sessions card, and Recent list now include Claude Cowork sessions alongside Claude Code, with per-day cost attribution: a Cowork task spanning midnight contributes only today's spend to Today. Cowork rows show "Cowork" in place of a project path; the Projects stat stays CLI-only because a Cowork workspace ID is an opaque UUID, not a project directory.
- **Data-coverage badge.** The Analytics header now states that the estimate covers only the transcripts still present on disk, flagging deleted sessions (history.jsonl vs on-disk) and the settings.json `cleanupPeriodDays` retention window.
- **Skill and command tool-restriction linting.** New SKL013 and CMD007 rules flag malformed or self-contradictory `allowed-tools` / `disallowed-tools` frontmatter (a tool listed as both allowed and disallowed, or an unknown tool name) in skills and commands.
- **AutoMode governance linting.** CFG008 flags `allowAllClaudeAiMcps`, which trusts every claude.ai MCP server without per-server review; HRD012 flags an `autoMode` block that has no `hard_deny` baseline, the non-bypassable stops that keep an unattended run from taking destructive actions on its own.

### Improvements

- **Session titles.** Sessions now display a human-readable title: a `/rename` custom title wins, then the slug, then the first user message (truncated to 80 characters), falling back to the session ID prefix instead of showing a bare ID.
- **Fast-mode billing.** Sessions run in Claude Code fast mode are now billed at the correct 2x rate, across both Claude Code and Cowork.
- **Settings rail.** Surfaces `fallbackModel` (a string or up-to-three-model array) in the Model section, plus friendly labels for `disableBundledSkills` and `disableSkillShellExecution`.
- **Hooks rail.** Recognizes the `terminalSequence` field (Claude Code 2.1.141) and surfaces the `MessageDisplay` event (2.1.152), and adds icons for the SessionEnd, SubagentStop, PreCompact, PostToolUseFailure, and FileChanged events.

### Bug Fixes

- Orphan streams (aborted, with no `stop_reason` record anywhere in the file) now bill the maximum cumulative-usage record per `message.id` instead of the first, smallest intermediate, so aborted-stream cost matches what Anthropic charged.
- Context-forking subagent files (`agent-acompact-*`, `agent-aside_question-*`) replay the parent transcript verbatim; their copied `message.id`s are now excluded from totals, recovering ~$31 of phantom spend on a heavy-subagent machine. Regular `agent-<hash>` subagent files are unaffected.
- Resumed-session cost is now attributed to the day it was incurred. Each session carries a per-day breakdown keyed by local calendar day, and analytics aggregate only the in-range days, so a `/resume` of an older session no longer adds its earlier-day cost to today's total. Every analytics card and the popover Today stat reconcile under any time range, and a drop in the daily Models chart is fixed.

## Claudoscope 0.7.0

*May 12, 2026*

### New Features

- **Cowork rail.** New rail that surfaces Claude Cowork sessions (the agentic mode in the Claude desktop app) alongside your Claude Code data. Reads from `~/Library/Application Support/Claude/` and renders session metadata, per-model token and cost stats, generated files, and the full transcript in the same chat view used for Claude Code sessions. The rail appears automatically when Cowork is configured and at least one session exists; stays hidden otherwise.
- **Cowork spend in Analytics.** The Est. Cost card headline now shows the combined CLI + Cowork total, with an inline "+ $X.XX Cowork" subtitle mirroring the existing "+ X cache" pattern on Tokens. Hidden when an analytics project filter is active because Cowork's project namespace does not overlap with CLI projects.
- **Hardening rail and 1-click baseline installer.** New rail that installs a vendor-neutral security baseline into `~/.claude/` across five layers: permissions deny rules and sandbox isolation, seven PreToolUse and PostToolUse shell hooks (credential scanner, command validator, public-repo push guard, proprietary-file flag, 14-day package-age check, `git reset --hard` intent guard, post-tool credential scan), AutoMode soft-deny rules requiring explicit user intent for high-risk operations, a marker-wrapped governance block in `CLAUDE.md`, and a security-awareness skill the agent can consult on demand. The installer takes a full backup of `~/.claude/` before writing anything; Install, Reinstall, Revert, and Uninstall are always one click away.
- **Hardening drift detection.** Eleven HRD lint checks (HRD001 through HRD011) verify each layer stays in place after install, grouped by layer in the rail with one-click fixes for the drift cases that can be auto-repaired. Per-rule REMEDIATION cards show entry-specific rationale for each deny rule and hook script in plain language, so non-security specialists can read why a given block matters.
- **Trusted Sources sheet.** Curate the hosts, package registries, and repositories the agent can reach without explicit approval. Each entry shows a plain-language description of what it unlocks, so the security implications are clear without reading external docs.

### Improvements

- Cowork rail icon uses the checklist SF Symbol so it reads as task-tracking at a glance.
- README rewritten with a top-level Hardening section and a Cowork rail subsection; intro and TOC updated to reflect the two new pillars.

## Claudoscope 0.6.2

*April 30, 2026*

### New Features

- Settings rail surfaces the new `prUrlTemplate` top-level key from Claude Code 2.1.119, rendered in the Attribution section alongside the commit and PR templates.
- Settings rail gains a Themes section that enumerates `~/.claude/themes/*.json` (introduced in Claude Code 2.1.118), showing each theme by name with its modification date and an "active" badge for the theme referenced in `~/.claude.json`.
- Hooks rail detail view links to the official Anthropic hooks documentation and notes that PostToolUse and PostToolUseFailure hook stdin includes `duration_ms` as of Claude Code 2.1.119.
- Secret scanning: new SEC009 detector for Slack incoming webhook URLs (ERROR severity, supports services/workflows/triggers paths).
- Secret scanning: new SEC010 critical-credential tier (ERROR severity) for account-level platform tokens, covering Stripe live/prod keys, Stripe webhook signing secrets, OpenAI service-account and admin keys, Anthropic admin keys, Azure storage AccountKeys, and Vault tokens. The latter four were previously WARNING under SEC007.

### Improvements

- Cold-cache initial scan is 34.9% faster: `SessionParser` collapses the prior two-disk-pass design into a single streaming pass with in-memory replay. Measured on a 2,052-session install: 37.3s → 24.3s cold (median, n=5), ~39s → ~26s warm.
- Analytics charts on 30-day and all-time ranges no longer render garbled X-axis labels. A new `stridedDateXAxis` modifier picks ~7 evenly spaced dates (always keeping the last) and is shared across the I/O, cache, hit-ratio, model-cost, and effort charts.
- File watcher now treats edits under `~/.claude/themes/` as config changes, so the new Themes section live-reloads on edit.
- SEC007 platform-token detection extended with verified vendor formats from gitleaks: GitHub OAuth/server/user/refresh tokens (gho_/ghs_/ghu_/ghr_), OpenAI project keys (sk-proj- with the T3BlbkFJ literal anchor), legacy OpenAI keys (\bsk-…{48}\b with word boundaries), SendGrid (SG.x.y), Shopify (shp[atspc]_ all four prefixes), DigitalOcean (dop_v1_), Linear (lin_api_), and PyPI (pypi-AgEIcHlwaS5vcmcC… macaroons).
- SEC004 keyword group extended to recognize `aws_secret_access_key` and `aws_secret_key` assignments.
- SEC007 sk-ant- alternation now uses a negative lookahead to avoid double-matching admin01 keys (which live in SEC010).

### Bug Fixes

- Fix cost overcount on tool-heavy sessions where Claude Code re-persists the same Anthropic API response across tool-use turn boundaries. Each copy shared `message.id` but carried a unique `uuid`, so the prior uuid-only dedup missed them. Now deduped by `message.id` with uuid as fallback. One observed case: $1616 → $898 (matching the actual Anthropic bill).
- Recover hidden cost from subagent records and aborted streams. Subagent JSONL files share the parent's `sessionId` and were being skipped by the continuation parent-skip branch (~14% of cost dropped on heavy-subagent setups); now detected via per-record `isSidechain` with a path-based fallback. Aborted streams produce orphan `message.id`s with no `stop_reason` record anywhere in the file; a first pass collects stop-reason ids, the second pass bills records carrying either `stop_reason` or an orphan id. Combined, the gap from a real Vertex bill closes from ~10% to <5%.
- Fix existing GitHub token regex (`ghp_[A-Za-z0-9_]{36}`) which incorrectly allowed underscores in the token body. Per GitHub's published spec, token bodies are base62 only.

## Claudoscope 0.6.1

*April 19, 2026*

### New Features

- Hooks rail now merges rules from all five sources (user, project, project-local, plugin, managed) with a SOURCE label per rule. Previously only ~/.claude/settings.json was read, silently hiding hooks shipped by plugins or defined per-project.
- Hook events beyond the legacy whitelist (SessionEnd, PostToolUseFailure, PreCompact, FileChanged, etc.) now surface automatically.

### Improvements

- Startup no longer saturates CPU on large session directories: streaming JSONL reader, lightweight metadata-only decode pass, bounded parallel parsing (cap 8), and cooperative cancellation. Scan progress banner during initial load.
- Config live-reload: edits to ~/.claude/settings.json and the plugin cache now reflect without app restart (debounced 250ms pipeline).
- Plugin, command, and skill version selection switched from lex sort to mtime, so 1.10.0 correctly beats 1.9.0 and timestamped builds beat "unknown".
- Multi-day sessions split across UTC days proportionally by elapsed seconds; tier costs computed per-session so the breakdown reconciles with actualCost.
- Unrecognized models no longer silently priced as Sonnet; analytics skip them via an isUnknown sentinel.

### Bug Fixes

- Fix EXC_BAD_ACCESS crash from concurrent dictionary mutation in delta reads (added @MainActor isolation, NSLock-protected DeltaTracker).
- Fix FSEvents callback use-after-free during teardown via a StreamBox weak-reference pattern.
- Recover from FSEvents overflow (MustScanSubDirs, KernelDropped, UserDropped) instead of silently losing events.
- Fix sidebar "COST BY PROJECT" stuck on a hardcoded 30-day window while displaying the user-selected time range label.
- Fix silent parse errors making projects disappear from the UI; failures now log to Console.app.
- Fix ISO8601 timestamp parsing falling back inconsistently across four call sites.
- Fix sluggishness from @SceneStorage in views hosted outside the SwiftUI scene lifecycle (replaced with @AppStorage).
- Fix Custom date pickers being clipped in the TimeRangePicker header.
- Fix per-session view state leaking between sessions in Tools and Agent Tree panels.
- Fix update-check tasks orphaning on popover dismissal.
- Fix cache-tier attribution treating present-but-empty breakdown as authoritative; legacy total now wins, attributed to the 5m tier.
- Fix tool-result dedup ordering (now first-write-wins).
- Fix AnyCodable mis-decoding numeric 1/0 as Bool.

## Claudoscope 0.6.0

*April 13, 2026*

### New Features

- Claude Code v2.1.90+ support: recognize Monitor, EnterWorktree, ExitWorktree tools with proper icons and exec-category classification
- PermissionDenied hook event displayed in Hooks rail
- Git worktree badge: sessions that use worktree tools show a cyan branch icon in the sidebar
- CFG007 lint rule: flags when skill shell execution is unrestricted (scoped to users with plugins)
- Pre-release security audit pipeline: scans npm and Swift/SPM dependencies against GHSA and OSV databases, gating releases on HIGH+ CVEs
- MDM-managed auto-update preference: respects macOS Configuration Profiles, disables toggle with "Managed by your organization" label

### Improvements

- disableSkillShellExecution status surfaced in Settings > Security
- Auto-update reworked: async process execution, .bak rollback safety, improved cleanup on failure or cancellation
- Secret scanning hardened with new patterns, real-time tail scanning fixes, and unit tests
- Secret alert deduplication and detection UX improvements
- Download counter validates release asset existence before counting
- README updated: 45 lint rules, all hook event types, session badges, CFG checks documented

### Bug Fixes

- Fix wrangler 3.x CVEs by upgrading to 4.81.0

## Claudoscope 0.5.0

*March 31, 2026*

### New Features

- 9 observability features: turn duration analytics (histogram + percentiles), effort level classification with donut chart and cost breakdown, subagent tree visualization in session detail, error pattern detection with sidebar badges, idle/zombie session detection, config health linter expansion (CFG001-CFG006), parallel tool call badges, SEC008 ENV_SCRUB correlation
- Timeline overhaul: time-gutter layout with project color strips, adaptive gap spacing, message type differentiation, collapsed project badges, clickable session navigation
- Config health auto-fix: one-click Apply Fix for CFG006 (subprocess env scrub)
- Rich markdown rendering in plans detail panel
- Resizable sidebar (180-400pt) with persistence and double-click reset
- Tooltips on truncated project names in sessions and analytics sidebars

### Improvements

- Wider rail buttons for better label readability
- Secret alerts fire immediately via App.init() callback, no longer require popover to be open; alert panel centered on screen

### Bug Fixes

- Fix crash in cache analytics when all sessions have zero cache tokens
- Fix UUID dedup bug (scoped per parseMetadata call) and streaming intermediate filtering

## Claudoscope 0.4.7

*March 21, 2026*

- Tabbed analytics view with Overview, Cache, and Models segments
- Actionable cache dashboard: busting detection, stability callout, 5m/1h TTL tier breakdown, per-session efficiency table, model-aware savings
- Model analysis tab: daily cost by model chart, model efficiency table, what-if Opus-to-Sonnet calculator
- Tools rail: per-session tool call extraction, category breakdown (Read/Write/Exec/Other), tool analytics
- Command palette (Cmd+K) for quick navigation between rails
- Subagent session content loading and badge for secret scan findings
- Replaced NSPanel update dialogs with native SwiftUI Window scenes
- Improved health check scoring and popover UX

## Claudoscope 0.4.6

*March 17, 2026*

- Added Config Health screen: 19 lint rules across CLAUDE.md, rules, and skills with group-by-rule view, severity filters, health gauge, and one-click rescan
- Added session health checks (SES001-SES004) surfacing expensive, long, or idle sessions
- Added secret detection scanning session files for leaked credentials with entropy filtering, context lines, and reveal toggle
- Added real-time secret alerts with settings toggle
- Redesigned Config Health with category navigation and human-readable rule names
- Added What's New button and full release notes in Settings > Updates
- Replaced loading skeleton with animated logo in menu bar popover
- Improved typography: bumped scale +1pt across all views
- Refactored 9 monolithic source files into ~40 focused modules

## Claudoscope 0.4.5

*March 16, 2026*

- Added "Skip This Version" option to update popup
- "Later" now clears badge and re-prompts on next check cycle
- Fixed update popup showing twice on manual "Check for Updates"
- Fixed Dock icon disappearing when dismissing update popup while main window is open
- Fixed download cancel button not working
- Fixed URLSession leak during update downloads
- Skip redundant update check on launch if checked within the last hour

## Claudoscope 0.4.4

- Support tracking multiple active sessions simultaneously in the menu bar popover
- Active sessions display in a unified card with compact rows and a pulsing indicator
- Scrollable active sessions section when more than 4 sessions are running

## Claudoscope 0.4.3

*March 16, 2026*

- Fixed release notes text not rendering in update and What's New popups
- Improved auto-update relaunch to avoid overlapping processes

## Claudoscope 0.4.0

*March 16, 2026*

- (Yanked, fixes were incorrect)

## Claudoscope 0.3.9

*March 16, 2026*

- Added bundled changelog for reliable "What's New" popup after updates
- Added download count badge to README
- Added changelog gate to release script
- Fixed today's sessions filter to use proper date comparison
- Fixed watcher re-parse UUID deduplication reset
- Fixed project ID derivation for subagent paths

## Claudoscope 0.3.8

*March 16, 2026*

- Added download tracking for Homebrew installs
- Fixed Dock icon not appearing when opening Dashboard
- Fixed version not updating on auto-update
- Fixed phantom sonnet in Model Distribution chart

## Claudoscope 0.3.7

*March 15, 2026*

- Maintenance release with internal improvements

## Claudoscope 0.3.6

*March 15, 2026*

- Added project-scoped memory rail
- Fixed path decoding issues

## Claudoscope 0.3.5

*March 15, 2026*

- Added "What's New" dialog after auto-updates
- Added update notification popups

## Claudoscope 0.3.4

*March 15, 2026*

- Improved cost estimation accuracy
- Fixed streaming record deduplication
- Added subagent session scanning

## Claudoscope 0.3.3

*March 15, 2026*

- Added automatic self-updating via GitHub Releases
- Fixed cost estimation: per-message pricing, cache write tiers, UUID dedup

## Claudoscope 0.3.2

*March 15, 2026*

- Added MCP server loading from project-level .mcp.json
- Added onboarding popup and about overlay
- Switched to Anthropic pricing as default
- Fixed pricing table alignment

## Claudoscope 0.3.1

*March 15, 2026*

- New app logo
- Added Homebrew cask distribution
- Added release automation

## Claudoscope 0.3.0

*March 15, 2026*

- Added Settings view with Security, Attribution, Plugins, and Account sections
- Added rich markdown rendering for skills and commands
- Replaced MCP list with expandable card grid