- Remove legacy .fusion/memory.md fallback references and normalize prompts/docs to .fusion/memory/MEMORY.md
- Stop legacy mirror writes and fallback reads in core memory backend and project memory flows
- Update engine worktree boundary checks and tests for canonical memory file handling
- Align dashboard memory/settings surfaces and route tests with canonical memory behavior
- Add model-favorites persistence test coverage for mission interview and new agent dialogs
- Add structured warning logs for plugin-runner fallback paths instead of silent catches
- Add warning telemetry in agent-tools for agent-memory directory, QMD search, and index refresh failures
- Log cron-runner session disposal failures and update agent-logger persistence warnings to use logger output
- Expand engine tests to cover warning behavior across plugin-runner, agent-tools, cron-runner, and agent-logger
- Update core and engine prompt text to reference .fusion/memory/ with MEMORY.md and daily-note guidance
- Allow worktree tool boundary access to .fusion/memory/ files while preserving legacy .fusion/memory.md compatibility
- Revise memory compaction/insight extraction messaging and related type/docs comments to use canonical memory paths
- Refresh dashboard and test expectations across core/engine/dashboard/docs for the new memory path wording
Anthropic now requires `anthropic-beta: oauth-2025-04-20` on /api/oauth/usage
for OAuth-scoped tokens; without it the endpoint returns 401 "OAuth
authentication is currently not supported" and the dashboard falls back to
a PTY-based CLI parser that times out at 75s. Mirrors the header the Claude
CLI (2.1.114) sends from `claude /usage`, and also aligns the refresh call
with the CLI's JSON body + `scope` field so token refresh keeps working.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- Add HEARTBEAT_NO_TASK_SYSTEM_PROMPT for no-task heartbeat sessions with only non-task-scoped tooling guidance
- Switch heartbeat session setup to choose the no-task base prompt when executing without task context
- Keep agent instruction and memory enrichment on top of the selected base prompt for both task and no-task runs
- Extend heartbeat tests to verify no-task prompts exclude task-scoped tool references and are passed into createKbAgent
- Follow symlink entries that resolve to directories so linked markdown trees are discovered
- Add guarded scan diagnostics for non-ENOENT directory read failures without noisy missing-dir logs
- Expand markdown scanner regression tests for symlinks, empty roots, deep nesting, and .md-like directory names
- Clarify the Documents view empty-state copy to say no Markdown files were found in the project
- Trigger onError for unexpected AbortError failures in streamChatResponse while ignoring user-initiated closes
- Keep quick chat user messages on send errors, reset streaming state, and reload messages from the server
- Persist partial assistant output (including thinking-only fallbacks) when AI processing fails mid-stream
- Add regression coverage for stream abort handling, quick chat error UX, and partial-response persistence
- Replace outdated dashboard screenshots across the guide and demo asset
- Add new view screenshots for chat, documents, list, mailbox, memory, nodes, roadmaps, and skills
- Update docs/dashboard-guide.md to reference the new visuals for current UI flows
- Add fn skills quick-start examples to the README command snippets
- Add a dedicated fn skills section in the CLI reference with search/install usage and option docs
- Document --interactive support for daemon and desktop command usage and option tables
- Expand settings export/import flag documentation and refresh the option cross-reference matrix
Dropping the render-phase setTasks([]) on projectId change avoids the
blank flash and full empty→populated Board reconcile that made project
switches feel like a multi-second hang. The existing fetch/version
guards already reject late responses and stale SSE events, so the
previous project's rows safely stay on screen until the new fetch lands.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- Log final-path git reset --merge cleanup failures in aiMergeTask with task-scoped warning details
- Preserve the original "all 3 attempts exhausted" merge error even when cleanup reset fails
- Add merger retry tests covering successful final cleanup and failed final cleanup logging behavior
- Add a patch changeset for @gsxdsm/fusion documenting the improved merger diagnostics
- Remove the agents tab button and showAgentsTab prop from MobileNavBar.
- Stop passing showAgentsTab from App to keep mobile navigation aligned with current feature exposure.
- Update MobileNavBar and mobile regression tests to remove agents-tab assertions and verify agents access via header fallback when mobile nav is disabled.
- Move the core DocumentsView, document group, and document card styles to global scope so they apply on desktop and tablet
- Keep the max-width: 768px block focused on mobile-only agent and skills layout overrides
- Preserve existing design tokens and interaction states while restoring expected non-mobile DocumentsView styling
- Extend useMemoryData with multi-file memory documents, settings state, and retrieval test helpers
- Upgrade MemoryView working memory tab to manage multiple files with file switching and editor actions
- Add configuration controls for dreams mode and auto-summarize behavior in memory settings
- Integrate qmd retrieval testing in the engines tab and surface retrieval results in dedicated cards
- Add styles for memory settings panels and retrieval result cards to match dashboard UI patterns
- Remove the trailing-newline guard when flushing pending SSE buffer data in streamChatResponse
- Re-enable the regression test covering final done events without a trailing blank line
- Add parser coverage for chunk-split events, newline-terminated streams, and incomplete trailing fragments
- Add a patch changeset for @gsxdsm/fusion documenting the SSE final-buffer flush fix
- Add backend Markdown scanner and /project-files/md endpoint with directory exclusions, depth/size limits, and search filtering
- Extend dashboard API and useDocuments hook to fetch task documents and project markdown files in parallel with resilient partial-failure handling
- Update Documents view with a collapsible Project Files section, inline metadata previews, and on-demand file content expansion from workspace API
- Add and refresh hook and file-service tests to cover markdown scanning, query propagation, project scoping, and fetch error behavior
- Add token-based styles for project file cards and responsive mobile adjustments in the documents panel
- Normalize GET /api/skills/:id/content handling by decoding encoded IDs and returning consistent invalid-id and not-found responses
- Update readSkillContent to use stat-based file vs directory detection, read SKILL.md robustly, and list non-SKILL.md entries
- Refine SkillsView interactions to ignore toggle clicks, support explicit retry of failed content fetches, and show a SKILL.md fallback placeholder
- Add/adjust dashboard route, component, and mobile CSS tests to cover content route behavior, retry flow, and touch-friendly detail panel styles
- Convert seed-sample-nodes into a test-only helper and remove direct execution against the real central database
- Replace private-looking host examples with RFC 5737 test-net addresses in node connection docs and ConnectNodeModal expectations
- Add Vitest aliases for @fusion/core and @fusion/engine to stabilize cross-package test imports
- Load AgentReflectionService via local engine source path in in-process runtime to avoid alias resolution failures
- Replace emoji role icons in AgentListModal, AgentsView, and NewAgentDialog with consistent geometric symbols
- Update agent preset catalog icons, including reviewer and template agent presets, to professional symbol-based glyphs
- Change the default unknown-role icon fallback from a robot emoji to a neutral symbol
- Update NewAgentDialog and agent-presets tests to assert symbol icons and broaden icon validation beyond emoji ranges
- Add defensive cleanup for in-memory task tracking when tasks move state or agents pause
- Improve stuck-detection and self-healing flow to reduce leaked state and missed recovery paths
- Add logging around previously swallowed errors and tighten executor cleanup behavior
- Expand restart and self-healing reliability tests to cover regression scenarios
- Recompute slice status when adding a feature so stale "complete" slices cascade back to milestone and mission statuses
- Require checkMissionCompletion to verify every feature is done before marking a mission complete
- Add mission-wide status-chain recomputation when stale complete slices or milestones are detected during reconciliation
- Expand core and engine tests to cover addFeature downgrades, stale complete states, and feature-level completion checks
- Add memoryView to KNOWN_EXPERIMENTAL_FEATURES and derive memoryEnabled in useAppSettings
- Gate Memory view rendering and board fallback in App based on memoryEnabled
- Hide Memory actions in Header overflow and MobileNav more sheet unless experimentalFeatures.memoryView is enabled
- Extend App, Header, and MobileNav tests to cover memoryView enabled/disabled behavior
- Extract generateFeatureSuggestionsCore() that accepts onText/onThinking callbacks
- Refactor generateFeatureSuggestions() to use core with no-op callbacks
- Add generateFeatureSuggestionsStream() that uses core with SSE callbacks
- Add SseCallbacks interface for streaming output types
