package.json (1)

85-85: Remove @types/uuid from runtime dependencies.

uuid already ships its own TypeScript declarations, so keeping @types/uuid here is redundant, and putting it in dependencies ships a dev-only package in production. Please drop it, or at minimum move it to devDependencies.

Also applies to: 145-145

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@package.json` at line 85, The package.json currently lists "@types/uuid":
"^10.0.0" as a runtime dependency; remove that entry (or move it to
devDependencies) because the uuid package includes its own TypeScript types and
`@types/uuid` should not be shipped in production; update the dependencies block
to delete the "@types/uuid" line (or add it under devDependencies if you need it
for local tooling) and run a quick npm/yarn install to verify the lockfile
updates.

pkg/baseds/baseds.go (1)

24-29: Make BadgeEvent actions mutually exclusive.

Clear, ClearAll, ClearById, and Badge can all be set at once with this shape, so malformed events have no explicit precedence rule. A tagged-union style payload or a small validation helper would make this shared contract much harder to misuse.

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@pkg/baseds/baseds.go` around lines 24 - 29, BadgeEvent currently allows
multiple action fields to be set at once; add a validation helper (e.g., func (e
*BadgeEvent) Validate() error or ValidateBadgeEvent(e *BadgeEvent) error) that
enforces a tagged-union rule: exactly one of Clear, ClearAll, ClearById
(non-empty string), or Badge (non-nil) must be set, returning a clear error when
zero or >1 are set; call this validator wherever events are unmarshaled or
processed so malformed payloads are rejected early and make BadgeEvent's
contract explicit.

eslint.config.js (1)

79-81: Scope the get/e exemptions more narrowly.

These patterns suppress unused-variable warnings for any parameter named get and any local/caught variable named e across the repo, which will hide real dead-code regressions outside the intended Jotai/error-handler cases. Prefer _get / _e, or move this override to the specific files that need it.

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@eslint.config.js` around lines 79 - 81, The current global ignore regexes
(argsIgnorePattern, varsIgnorePattern, caughtErrorsIgnorePattern) are too
broad—narrow them to only match underscored placeholders (e.g., change matches
to require a leading underscore like `_get` and `_e`) or remove the global
`get`/`e` exemptions and relocate the relaxed rules to the specific files or
overrides that need them; update the regexes for argsIgnorePattern,
varsIgnorePattern and caughtErrorsIgnorePattern accordingly or add an ESLint
override block scoped to the specific file patterns that require these
exceptions.

pkg/util/unixutil/unixutil_unix.go (1)

72-78: Consider handling EPERM for processes owned by other users.

The current implementation returns false when syscall.Kill(pid, 0) returns an error, but EPERM indicates the process exists but the caller lacks permission. For the badge watch use case (monitoring a PID the caller started), this works correctly since the caller owns the process.

If broader use is anticipated, consider:

🔧 Optional fix to handle EPERM

 func IsPidRunning(pid int) bool {
 	if pid <= 0 {
 		return false
 	}
 	err := syscall.Kill(pid, 0)
-	return err == nil
+	return err == nil || err == syscall.EPERM
 }

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@pkg/util/unixutil/unixutil_unix.go` around lines 72 - 78, The IsPidRunning
function currently treats any error from syscall.Kill(pid, 0) as “not running”;
update it to treat EPERM as success (process exists but not permitted) by
checking the returned error against syscall.EPERM (or errors.Is(err,
syscall.EPERM)) and returning true in that case, while still returning false for
other errors; adjust IsPidRunning to call syscall.Kill(pid, 0), then if err ==
nil return true, else if the error is EPERM return true, otherwise return false.

frontend/app/tab/tabbar-model.ts (1)

4-15: Consider removing the empty singleton class.

TabBarModel is now a hollow singleton with no state or methods beyond the pattern itself, and it is not used anywhere in the codebase. If no longer needed, removing it would eliminate dead code.

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@frontend/app/tab/tabbar-model.ts` around lines 4 - 15, TabBarModel is an
empty, unused singleton (class TabBarModel with private static instance and
getInstance) that constitutes dead code; remove the entire TabBarModel class
declaration and any associated references (instance, getInstance) from the
codebase, or if it must be preserved for future use, add meaningful
state/methods and use it where intended—preferably delete the class to eliminate
dead code and run the test suite/TS compiler to ensure no imports/reference
errors remain.

frontend/preview/previews/tab.preview.tsx (1)

24-28: Duplicate badge IDs in preview data.

Lines 26-27 both use badgeid: "b1" for different badges within the same tab. While this is preview/demo data, using unique IDs would better represent production behavior and avoid potential confusion when debugging.

Suggested fix

         badges: [
             { badgeid: "b2", icon: "circle-check", color: "#4ade80", priority: 3 },
-            { badgeid: "b1", icon: "circle-small", color: "#fbbf24", priority: 1 },
-            { badgeid: "b1", icon: "circle-small", color: "red", priority: 1 },
+            { badgeid: "b1a", icon: "circle-small", color: "#fbbf24", priority: 1 },
+            { badgeid: "b1b", icon: "circle-small", color: "red", priority: 1 },
         ],

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@frontend/preview/previews/tab.preview.tsx` around lines 24 - 28, The preview
data for badges contains duplicate badge IDs ("b1") in the badges array inside
tab.preview (badges: [...] entries); update the badgeid values so each object in
the badges array has a unique identifier (e.g., change the second "b1" to a new
unique id) to mirror production behavior and avoid ID collisions when rendering
or debugging the preview.

frontend/app/store/badge.ts (1)

194-210: Different secondary sort directions between sortBadges and sortBadgesForTab.

sortBadges sorts by badgeid descending (b.badgeid < a.badgeid), while sortBadgesForTab sorts ascending (a.badgeid < b.badgeid). If this is intentional (e.g., newest first for blocks, oldest first for tabs), consider adding a comment explaining the distinction.

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@frontend/app/store/badge.ts` around lines 194 - 210, The two comparators in
sortBadges and sortBadgesForTab disagree on the secondary sort direction
(badgeid): sortBadges currently orders badgeid descending while sortBadgesForTab
orders ascending; decide whether they should match or the difference is
intentional and then implement that change: either make both functions use the
same comparator logic for badgeid (update sortBadges or sortBadgesForTab to use
the same a.badgeid < b.badgeid ? -1 : ... pattern) or, if the asymmetry is
intentional, add a short clarifying comment above sortBadges and
sortBadgesForTab explaining the rationale (e.g., "blocks: newest first vs tabs:
oldest first") so future readers understand the differing secondary sort
directions.

frontend/app/app.tsx (1)

253-268: Tab badge clearing lacks staleness check.

Unlike the block clearing logic (lines 245-248), the tab clearing effect doesn't verify the tab is still focused before clearing. Consider adding a similar check.

Suggested improvement

         const timeoutId = setTimeout(() => {
             if (!document.hasFocus()) {
                 return;
             }
+            const currentTabId = globalStore.get(atoms.staticTabId);
+            if (currentTabId !== tabId) {
+                return;
+            }
             clearBadgesForTabOnFocus(tabId);
         }, delay);

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@frontend/app/app.tsx` around lines 253 - 268, The tab-clear effect currently
clears badges unconditionally after the timeout; add a staleness/focus check
before calling clearBadgesForTabOnFocus(tabId): after the delay and before
invoking clearBadgesForTabOnFocus, re-check document.hasFocus() and verify the
tabId still corresponds to the active/focused tab (use the same focused-tab
check or ref used by the block-clearing logic, e.g., compare against the current
focusedTabId or ref the codebase uses) and only call
clearBadgesForTabOnFocus(tabId) when both checks pass.