stagent 0.10.0 → 0.11.0

package/docs/features/workflows.md

@@ -4,9 +4,9 @@ category: "feature-reference"
 section: "workflows"
 route: "/workflows"
 tags: [workflows, patterns, sequence, parallel, swarm, autonomous, templates, multi-step, blueprints, delays, drip]
-features: ["workflow-engine", "workflow-blueprints", "ai-assist-workflow-creation", "workflow-context-batching", "business-function-profiles", "workflow-step-delays"]
+features: ["workflow-engine", "workflow-blueprints", "ai-assist-workflow-creation", "workflow-context-batching", "business-function-profiles", "workflow-step-delays", "chat-conversation-templates", "workflow-run-history"]
 screengrabCount: 4
-lastUpdated: "2026-04-08"
+lastUpdated: "2026-04-15"
 ---
 
 # Workflows
@@ -48,6 +48,10 @@ The blueprint gallery offers pre-built workflow templates for common patterns:
 
 Click any blueprint to preview its steps, then create a workflow pre-populated with the template configuration. Customize steps, profiles, and runtimes to match your specific process.
 
+### Blueprints as Chat Templates
+
+Blueprints also double as **conversation templates** for the Chat surface. A blueprint can declare an optional `chatPrompt` field that provides a purpose-built opening prompt for chat use; if it is not set, the chat launcher falls back to `steps[0].promptTemplate`. This means the same blueprint asset powers three entry points — the Workflows gallery, the Chat empty-state picker, and the `/new-from-template` slash command — without requiring parallel definitions. See the [Chat](./chat.md) feature doc for the conversation-side flow.
+
 ### Tabs: All, Templates, Runs
 The workflow page organizes content into three tabs. "All" shows every workflow. "Templates" shows reusable patterns you can instantiate. "Runs" shows active and historical executions with their current status.
 

package/docs/index.md

@@ -1,7 +1,7 @@
 ---
 title: "Stagent Documentation"
 category: "index"
-lastUpdated: "2026-04-03"
+lastUpdated: "2026-04-15"
 ---
 
 # Stagent Documentation

package/docs/journeys/developer.md

@@ -6,7 +6,7 @@ difficulty: "advanced"
 estimatedTime: "30 minutes"
 sections: ["settings", "environment", "chat", "monitoring", "profiles", "workflows", "tables", "schedules", "delivery-channels"]
 tags: ["advanced", "developer", "settings", "environment", "cli", "api", "monitoring", "profiles", "ollama", "channels", "handoffs", "memory", "tables"]
-lastUpdated: "2026-04-08"
+lastUpdated: "2026-04-16"
 ---
 
 # Developer Guide
@@ -55,7 +55,7 @@ Riley sets up Ollama as the fifth runtime adapter for private, zero-cost executi
 
 Riley sets up Slack and Telegram as delivery channels for schedule notifications and bidirectional chat.
 
-![Settings data management section](../screengrabs/settings-data.png)
+![Settings Delivery Channels section](../screengrabs/settings-channels.png)
 
 1. Scroll to **Delivery Channels** in Settings
 2. Click **+ Add Channel** and select **Slack**
@@ -218,6 +218,29 @@ Riley builds the CLI for scripted operations and CI/CD integration.
 3. Test CRUD operations: `node dist/cli.js projects list`, `node dist/cli.js tasks create --title "CLI test"`
 4. Verify CLI-created entities appear in the web UI (shared SQLite database)
 
+### Step 14a: Understand the Runtime Capability Matrix
+
+Riley wants to know why some features show up on one runtime but not another. The answer lives in the runtime capability matrix.
+
+1. Open `src/lib/agents/runtime/catalog.ts` — each runtime adapter declares flags like `supportsSkillComposition`, `maxActiveSkills`, `hasNativeSkills`, `stagentInjectsSkills`, and `autoLoadsInstructions`
+2. The Chat Skills tab reads `supportsSkillComposition` to decide whether to enable multi-skill activation and "N of M active" reporting
+3. The SKILL.md injector reads `stagentInjectsSkills` to avoid duplicating context on runtimes (like Codex App Server and Claude Agent SDK) that load instructions natively
+4. When you wire a new feature that touches system prompts or skills, consult the matrix before deciding whether Stagent should inject something or trust the runtime to do so
+
+> **Tip:** The MCP task-tools boundary also validates the runtime ID now (runtime-validation-hardening). Malformed `runtimeId` values are rejected at the boundary with a clean error, rather than crashing the dispatcher.
+
+### Step 14b: Observe the Upgrade Detection Poller
+
+![Instance section showing dev mode status and upgrade detection gate](../screengrabs/settings-instance.png)
+
+1. Scroll to the **Instance** section in Settings to see the dev mode status and upgrade detection gate
+2. The hourly scheduler runs `git fetch` against the upstream remote and compares `HEAD` to `origin/main`
+3. When upstream is ahead, the sidebar shows an **Upgrade available** badge and the Instance card reports the commits-behind count
+4. Three consecutive poll failures escalate to a persistent inbox notification (three-strike dedup prevents notification floods)
+5. The upgrade session uses the `upgrade-assistant` profile, which allowlists **AskUserQuestion** — the assistant can ask direct questions mid-merge without hitting a generic permission prompt
+
+> **Tip:** The Instance card also shows bootstrap status. In dev mode (indicated by `STAGENT_DEV_MODE=true` or the `.git/stagent-dev-mode` sentinel), auto-upgrade machinery is skipped to avoid interfering with contributor workflows.
+
 ### Step 15: Verify Platform Health
 
 Riley performs a final platform health check.

package/docs/journeys/personal-use.md

@@ -6,7 +6,7 @@ difficulty: "beginner"
 estimatedTime: "30 minutes"
 sections: ["home-workspace", "dashboard-kanban", "projects", "chat", "tables", "schedules", "user-guide", "book"]
 tags: ["beginner", "solo", "tasks", "kanban", "chat", "tables", "schedules", "delivery-channels", "book"]
-lastUpdated: "2026-04-08"
+lastUpdated: "2026-04-16"
 ---
 
 # Personal Use Guide
@@ -69,9 +69,13 @@ Before setting up a formal project, Alex tries the Chat feature to brainstorm id
 2. Notice the **Tool Catalog** with a welcoming hero heading and suggested prompt categories (Explore, Create, Debug, Automate)
 3. Browse the **Smart Picks** row for personalized suggestions
 4. Type a question like "What pages should a developer portfolio website include?" and press Enter
-5. Review the AI response
+5. Review the AI response -- notice how Chat keeps a clean conversation thread, deduplicating similar messages so the history stays readable
+6. Try typing **@** in the composer to open the mentions popover -- you can reference tasks, projects, documents, profiles, and workflows inline to give the AI richer context
+7. Next time, try **Start from template** on the empty-state hero (or type `/new-from-template` in the composer) to open a conversation pre-primed with a blueprint like "Research a topic" or "Draft a plan"
 
-> **Tip:** Chat is perfect for quick brainstorming sessions. You do not need to create a project first -- just ask a question. The conversation history stays in the sidebar so you can return to it later.
+![@ mentions popover showing pinned entities, tasks, projects, and documents](../screengrabs/chat-mentions-popover.png)
+
+> **Tip:** Chat is perfect for quick brainstorming sessions. You do not need to create a project first -- just ask a question. The conversation history stays in the sidebar so you can return to it later. Pin useful filter + search combinations as **saved searches** (available from the `⌘K` palette under the **Saved** group) so you can jump back to them later.
 
 ### Step 5: Create a New Project
 
@@ -122,7 +126,10 @@ Alex creates the first task for the portfolio project.
 3. Write a **Description** with detail about requirements
 4. Assign the task to the **Portfolio Website** project
 5. Set **Priority** to High and leave **Status** as Planned
-6. Click **Create** to add the task to the board
+6. Open the **/** popover in the composer to browse available tools (create_task, execute_task, read_file) and entities -- these are the same capabilities the AI uses when running tasks
+7. Click **Create** to add the task to the board
+
+![/ popover Tools tab showing available runtime tools](../screengrabs/chat-tools-tab.png)
 
 > **Tip:** Write task descriptions as if you are briefing a colleague. The more specific you are, the better the AI agent results will be.
 
@@ -181,7 +188,7 @@ Alex wants Stagent to proactively check on the portfolio project every morning.
 
 Alex wants to receive schedule results on the go. Setting up a Telegram delivery channel takes less than two minutes.
 
-![Settings page showing provider and runtime configuration](../screengrabs/settings-auth.png)
+![Settings page showing Delivery Channels configuration](../screengrabs/settings-channels.png)
 
 1. Open **Settings** from the sidebar under **Configure**
 2. Scroll to the **Delivery Channels** section

package/docs/journeys/power-user.md

@@ -6,7 +6,7 @@ difficulty: "advanced"
 estimatedTime: "30 minutes"
 sections: ["dashboard-kanban", "profiles", "chat", "workflows", "tables", "schedules", "monitoring", "settings"]
 tags: ["advanced", "automation", "workflows", "profiles", "tables", "schedules", "monitoring", "bulk-operations", "ollama", "episodic-memory", "nlp-scheduling"]
-lastUpdated: "2026-04-03"
+lastUpdated: "2026-04-16"
 ---
 
 # Power User Guide
@@ -88,14 +88,16 @@ Sam wants to run privacy-sensitive tasks on local models with zero API cost.
 
 Sam creates schedules by describing them in plain English instead of writing cron expressions.
 
-![Schedules list with active and heartbeat schedules](../screengrabs/schedules-list.png)
+![New schedule form filled — Daily Standup Summary with weekday 9am cron](../screengrabs/schedules-create-form-filled.png)
 
 1. Navigate to **Schedules** and click **Create Schedule**
 2. In the interval field, type natural language: "every weekday at 10pm," "every 6 hours during business hours," or "twice daily at 9am and 5pm"
 3. A **preview** appears showing exactly how the system parsed the input
 4. Select **Heartbeat** type to add intelligence -- the agent evaluates a checklist before acting
 5. Add checklist items specific to your automation needs
-6. Save the schedule
+6. Save the schedule and click on it to open the **detail sheet** -- review firing history, next run time, and pause/resume controls
+
+![Schedule detail sheet with firing history and autonomous loop controls](../screengrabs/schedules-detail.png)
 
 > **Tip:** NLP scheduling removes the friction of cron syntax. If the preview does not match your intent, rephrase and it re-parses instantly.
 
@@ -103,10 +105,10 @@ Sam creates schedules by describing them in plain English instead of writing cro
 
 Sam customizes a workflow for a "Deploy & Verify" pipeline with specialized profiles at each step.
 
-![Workflows list showing existing workflow definitions](../screengrabs/workflows-list.png)
+![New workflow form with delay step — ML Pipeline Orchestrator with 3-step sequence](../screengrabs/workflows-create-form-delay.png)
 
 1. From the Workflows page, click **Create Workflow** (or customize a blueprint)
-2. Configure steps with different agent profiles at each stage
+2. Add steps with different agent profiles at each stage -- notice the **delay step** option for inserting timed pauses between steps (e.g., a 2-hour cooldown between deploy and verification)
 3. Use the business-function profiles for non-technical steps (e.g., Content Creator for documentation, Operations Coordinator for runbook updates)
 4. Save the workflow
 
@@ -146,7 +148,27 @@ Sam connects a table to the workflow engine so that row changes automatically ki
 
 > **Tip:** Table triggers turn spreadsheets into event-driven automation surfaces. Sam uses them to auto-run incident postmortems whenever an incident row status changes to "Resolved."
 
-### Step 11: Batch-Manage Tasks on the Kanban
+### Step 11: Create Tasks with AI Assist
+
+Sam uses AI Assist to turn a rough task idea into a fully specified, workflow-ready plan.
+
+![AI Assist returned — improved description, complexity analysis, and Swarm pattern recommendation](../screengrabs/dashboard-create-form-ai-assist.png)
+
+1. From the Dashboard, click **Create Task**
+2. Enter a rough title and description -- for example, "API Security Audit" with a one-line description
+3. Click **AI Assist** -- the system analyzes the task and returns an improved description, complexity rating, recommended execution pattern (e.g., Swarm), and a checkpoint flag
+4. Review the **breakdown** -- for complex tasks, AI Assist generates a multi-step workflow with per-step descriptions
+
+![AI Assist breakdown — 6-step security audit workflow with per-step descriptions](../screengrabs/dashboard-create-form-ai-breakdown.png)
+
+5. Click **Apply** to replace the original description with the AI-improved version, or **Convert to Workflow** to generate a full workflow from the breakdown
+6. The resulting workflow lands on the confirmation page, ready for review and customization
+
+![Task form after clicking Apply — improved description replaces original](../screengrabs/dashboard-create-form-ai-applied.png)
+
+> **Tip:** AI Assist is the bridge between "I have a vague idea" and "I have a production-ready plan." For complex tasks, it recommends the Swarm pattern where multiple agents tackle sub-steps in parallel -- this is how multi-agent swarm execution works in practice.
+
+### Step 12: Batch-Manage Tasks on the Kanban
 
 Sam cleans up the task board using bulk select mode.
 
@@ -160,7 +182,7 @@ Sam cleans up the task board using bulk select mode.
 
 > **Tip:** After a weekend of autonomous heartbeat runs, Sam's first Monday task is always a bulk cleanup.
 
-### Step 12: Schedule Automated Prompt Loops
+### Step 13: Schedule Automated Prompt Loops
 
 ![Schedules list showing active schedules and firing status](../screengrabs/schedules-list.png)
 
@@ -170,7 +192,7 @@ Sam cleans up the task board using bulk select mode.
 4. Verify stop conditions and delivery channels
 5. Toggle **Pause/Resume** as needed
 
-### Step 13: Leverage Episodic Memory
+### Step 14: Leverage Episodic Memory
 
 Sam notices agents are re-researching the same topics. Episodic memory lets agents retain factual knowledge across executions.
 
@@ -184,7 +206,7 @@ Sam notices agents are re-researching the same topics. Episodic memory lets agen
 
 > **Tip:** Episodic memory means a Financial Analyst profile that researches a company once can recall that research in future tasks without re-doing the work. It builds institutional knowledge automatically.
 
-### Step 14: Watch Agent Execution in Real-Time
+### Step 15: Watch Agent Execution in Real-Time
 
 ![Agent monitoring dashboard showing real-time execution logs](../screengrabs/monitor-list.png)
 
@@ -196,16 +218,25 @@ Sam notices agents are re-researching the same topics. Episodic memory lets agen
 
 > **Tip:** The Monitor is Sam's operational dashboard. When something goes wrong in an autonomous loop at 3am, the execution traces are the fastest path to diagnosis.
 
-### Step 15: Use Chat Suggested Prompts
+### Step 16: Use Chat Suggested Prompts and Skill Composition
 
-![Chat interface with suggested prompt categories](../screengrabs/chat-list.png)
+![/ popover Skills tab with composition status indicator](../screengrabs/chat-skills-tab.png)
 
 1. Navigate to **Chat** and notice the **suggested prompt tabs** (Explore, Create, Debug, Automate)
 2. Click the **Create** tab to see prompts for creating tasks, workflows, and schedules
 3. Click a suggested prompt to populate the input
-4. Edit and send -- these prompts are optimized for the best agent responses
+4. Type **/** to open the slash popover -- browse four tabs: **Actions**, **Skills**, **Tools**, and **Entities**
+
+![/ popover with four tabs: Actions, Skills, Tools, Entities](../screengrabs/chat-slash-popover.png)
+
+5. Switch to the **Skills** tab — click **+ Add** to stack a second skill onto the conversation (e.g., a research skill plus a code-review skill running concurrently)
+6. If the new skill clashes with an active one, review the **conflict dialog** and decide which to keep — the engine's heuristic surfaces overlapping tools or contradictory instructions
+7. Watch the "N of M active" indicator; once you approach the runtime's `maxActiveSkills` cap, prompt-budget eviction automatically trims the oldest low-priority skill
+8. For a faster start, try **Start from template** or `/new-from-template` to open a conversation pre-primed from a workflow blueprint — the blueprint's optional `chatPrompt` provides the opening context
+
+> **Tip:** Skill composition only engages on runtimes where `supportsSkillComposition` is true. On other runtimes the Skills tab stays read-only with a hint, and the chat behaves as a single-skill session.
 
-### Step 16: Chain Everything Together
+### Step 17: Chain Everything Together
 
 Sam connects the dots: profiles define *how*, workflows define *what*, schedules define *when*, Ollama handles the *cheap stuff*, episodic memory provides the *knowledge*, and delivery channels deliver the *results*.
 
@@ -221,7 +252,7 @@ Sam connects the dots: profiles define *how*, workflows define *what*, schedules
 
 > **Tip:** Sam's automation philosophy: start small, observe, then expand. Run a workflow manually three times before scheduling it. Trust builds incrementally -- and so should autonomy.
 
-### Step 17: What's Next
+### Step 18: What's Next
 
 Sam's workspace is a fully autonomous operations engine. The next step is going deeper into the platform layer.
 

package/docs/journeys/work-use.md

@@ -6,7 +6,7 @@ difficulty: "intermediate"
 estimatedTime: "30 minutes"
 sections: ["projects", "chat", "documents", "tables", "workflows", "schedules", "cost-usage", "inbox-notifications", "delivery-channels"]
 tags: ["intermediate", "team", "documents", "tables", "workflows", "schedules", "costs", "inbox", "channels", "handoffs"]
-lastUpdated: "2026-04-03"
+lastUpdated: "2026-04-16"
 ---
 
 # Work Use Guide
@@ -56,6 +56,9 @@ Rather than clicking through every page, Jordan asks Chat for a quick status upd
 2. Type a question such as "What is the status of the Q2 planning project?"
 3. Review the response -- notice **Quick Access pills** linking to specific items
 4. Click a pill to jump directly to the referenced entity
+5. Use **@** mentions to pull specific entities into the conversation -- type `@` to browse tasks, projects, documents, and workflows without leaving the composer
+
+![@ mentions popover for referencing workspace entities in chat](../screengrabs/chat-mentions-popover.png)
 
 > **Tip:** Chat is the fastest way to get a cross-cutting status overview. Ask one question and follow the entity links.
 
@@ -107,6 +110,9 @@ Jordan has a CSV export from the previous quarter's performance review. Importin
 3. Review the column mapping -- Stagent auto-detects column types from the data
 4. Assign the table to the relevant project
 5. Click **Create** to import the data into a fully editable spreadsheet view
+6. Click into the table to open the **Data tab** with a spreadsheet-style editor -- edit cells inline, add rows, and review content
+
+![Table detail Data tab with spreadsheet editor and row content](../screengrabs/tables-detail.png)
 
 > **Tip:** Imported tables retain all their data as editable rows. You can add formula columns, charts, and workflow triggers on top of imported data.
 
@@ -150,7 +156,7 @@ Jordan schedules a weekly status summary using plain English instead of cron syn
 
 ### Step 11: Monitor Spending and Budgets
 
-![Cost and Usage dashboard showing spend metrics and budget gauges](../screengrabs/cost-usage-list.png)
+![Cost and Usage dashboard showing spend metrics and budget gauges](../screengrabs/costs-list.png)
 
 1. Click **Cost & Usage** in the sidebar
 2. Review **total spend** for the current billing period
@@ -188,12 +194,14 @@ Jordan's agents have been running in the background. Some have generated handoff
 2. For permission requests, review the tool name, arguments, and reason
 3. Choose **Approve**, **Deny**, or **Always Allow**
 4. For handoffs, approve to let the receiving agent begin work
+5. If an agent sends an **AskUserQuestion** notification, type your answer in the inline reply field — there is no approve/deny, just a short response that lets the agent resume
+6. Watch for **upgrade available** notifications and the occasional upgrade-failure alert (raised only after three consecutive poller failures, so these should be rare)
 
 ### Step 15: Configure Multi-Channel Notifications
 
 Jordan sets up Slack so the entire team receives schedule results and important alerts.
 
-![Settings page showing provider and runtime configuration](../screengrabs/settings-auth.png)
+![Settings page showing Delivery Channels configuration](../screengrabs/settings-channels.png)
 
 1. Open **Settings** and scroll to **Delivery Channels**
 2. Click **+ Add Channel** and select **Slack**
@@ -205,13 +213,14 @@ Jordan sets up Slack so the entire team receives schedule results and important
 
 ### Step 16: Manage Schedules
 
-![Schedules list with status indicators and next run timestamps](../screengrabs/schedules-list.png)
+![Schedule detail sheet showing firing history and pause/resume controls](../screengrabs/schedules-detail.png)
 
 1. Return to **Schedules**
-2. Locate the "Weekly Status Report" schedule
-3. Check **Last Run** and **Next Run** timestamps
-4. Click to review firing history and generated tasks
-5. Edit or pause as needed
+2. Locate the "Weekly Status Report" schedule and click to open its **detail sheet**
+3. Review the **firing history** with timestamps and outcomes -- suppressed heartbeat runs appear dimmed
+4. Check **Next Run** to confirm the cadence
+5. Use the **Pause/Resume** toggle when you need to temporarily silence a schedule during quiet periods
+6. Edit or delete as needed
 
 ### Step 17: What's Next
 

package/docs/manifest.json

@@ -1,6 +1,6 @@
 {
-  "generated": "2026-04-03T22:00:00Z",
-  "version": 3,
+  "generated": "2026-04-15T15:30:00Z",
+  "version": 4,
   "sections": [
     {
       "slug": "home-workspace",
@@ -29,7 +29,7 @@
       "path": "features/inbox-notifications.md",
       "route": "/inbox",
       "tags": ["inbox", "approvals", "notifications", "human-in-the-loop", "handoffs"],
-      "features": ["inbox-notifications", "ambient-approval-toast", "agent-async-handoffs"],
+      "features": ["inbox-notifications", "ambient-approval-toast", "agent-async-handoffs", "upgrade-detection", "upgrade-session"],
       "screengrabCount": 2
     },
     {
@@ -38,8 +38,8 @@
       "category": "feature-reference",
       "path": "features/chat.md",
       "route": "/chat",
-      "tags": ["chat", "conversations", "ai", "tool-catalog", "mentions", "model-selection", "channels"],
-      "features": ["chat-data-layer", "chat-engine", "chat-api-routes", "chat-ui-shell", "chat-message-rendering", "chat-input-composer", "bidirectional-channel-chat"],
+      "tags": ["chat", "conversations", "ai", "tool-catalog", "mentions", "model-selection", "channels", "skills", "templates", "filters", "saved-searches"],
+      "features": ["chat-data-layer", "chat-engine", "chat-api-routes", "chat-ui-shell", "chat-message-rendering", "chat-input-composer", "bidirectional-channel-chat", "chat-skill-composition", "chat-composition-ui-v1", "chat-conversation-templates", "chat-filter-namespace", "chat-pinned-saved-searches", "saved-search-polish-v1"],
       "screengrabCount": 4
     },
     {
@@ -59,7 +59,7 @@
       "path": "features/workflows.md",
       "route": "/workflows",
       "tags": ["workflows", "orchestration", "blueprints", "parallel", "swarm", "step-builder", "business-blueprints"],
-      "features": ["workflow-engine", "workflow-blueprints", "parallel-research-fork-join", "multi-agent-swarm", "autonomous-loop-execution", "workflow-ux-overhaul", "business-function-profiles"],
+      "features": ["workflow-engine", "workflow-blueprints", "parallel-research-fork-join", "multi-agent-swarm", "autonomous-loop-execution", "workflow-ux-overhaul", "business-function-profiles", "chat-conversation-templates", "workflow-run-history"],
       "screengrabCount": 3
     },
     {
@@ -69,7 +69,7 @@
       "path": "features/documents.md",
       "route": "/documents",
       "tags": ["documents", "upload", "preprocessing", "agent-context", "extraction"],
-      "features": ["document-manager", "file-attachment-data-layer", "document-preprocessing", "agent-document-context", "document-output-generation"],
+      "features": ["document-manager", "file-attachment-data-layer", "document-preprocessing", "agent-document-context", "document-output-generation", "chat-filter-namespace", "chat-pinned-saved-searches"],
       "screengrabCount": 2
     },
     {
@@ -129,7 +129,7 @@
       "path": "features/settings.md",
       "route": "/settings",
       "tags": ["settings", "authentication", "permissions", "presets", "budgets", "browser-tools", "ollama", "channels", "snapshots"],
-      "features": ["session-management", "tool-permission-persistence", "tool-permission-presets", "browser-use", "spend-budget-guardrails", "ollama-runtime-provider", "multi-channel-delivery", "bidirectional-channel-chat", "database-snapshot-backup"],
+      "features": ["session-management", "tool-permission-persistence", "tool-permission-presets", "browser-use", "spend-budget-guardrails", "ollama-runtime-provider", "multi-channel-delivery", "bidirectional-channel-chat", "database-snapshot-backup", "upgrade-detection", "upgrade-session"],
       "screengrabCount": 9
     },
     {
@@ -159,7 +159,7 @@
       "path": "features/provider-runtimes.md",
       "route": "cross-cutting",
       "tags": ["runtimes", "claude", "openai", "codex", "ollama", "providers", "architecture", "smart-router"],
-      "features": ["provider-runtime-abstraction", "openai-codex-app-server", "agent-integration", "ollama-runtime-provider", "anthropic-direct-runtime", "openai-direct-runtime", "smart-runtime-router"],
+      "features": ["provider-runtime-abstraction", "openai-codex-app-server", "agent-integration", "ollama-runtime-provider", "anthropic-direct-runtime", "openai-direct-runtime", "smart-runtime-router", "runtime-validation-hardening", "runtime-capability-matrix"],
       "screengrabCount": 2
     },
     {
@@ -169,7 +169,7 @@
       "path": "features/agent-intelligence.md",
       "route": "cross-cutting",
       "tags": ["ai", "routing", "loops", "swarm", "self-improvement", "ai-assist", "episodic-memory", "handoffs"],
-      "features": ["multi-agent-routing", "autonomous-loop-execution", "multi-agent-swarm", "agent-self-improvement", "workflow-context-batching", "ai-assist-workflow-creation", "agent-episodic-memory", "agent-async-handoffs"],
+      "features": ["multi-agent-routing", "autonomous-loop-execution", "multi-agent-swarm", "agent-self-improvement", "workflow-context-batching", "ai-assist-workflow-creation", "agent-episodic-memory", "agent-async-handoffs", "chat-skill-composition"],
       "screengrabCount": 0
     },
     {
@@ -179,7 +179,7 @@
       "path": "features/tool-permissions.md",
       "route": "cross-cutting",
       "tags": ["permissions", "approvals", "presets", "governance", "trust-tiers"],
-      "features": ["tool-permission-persistence", "tool-permission-presets", "ambient-approval-toast"],
+      "features": ["tool-permission-persistence", "tool-permission-presets", "ambient-approval-toast", "upgrade-session"],
       "screengrabCount": 1
     },
     {
@@ -199,7 +199,7 @@
       "path": "features/keyboard-navigation.md",
       "route": "cross-cutting",
       "tags": ["keyboard", "navigation", "command-palette", "accessibility", "shortcuts"],
-      "features": ["command-palette", "keyboard-shortcuts", "focus-management"],
+      "features": ["command-palette", "keyboard-shortcuts", "focus-management", "chat-conversation-templates", "chat-pinned-saved-searches", "saved-search-polish-v1"],
       "screengrabCount": 2
     },
     {
@@ -209,7 +209,7 @@
       "path": "features/shared-components.md",
       "route": "cross-cutting",
       "tags": ["components", "reusable", "page-shell", "detail-pane", "status-chip", "data-table"],
-      "features": ["page-shell", "detail-pane", "status-chip", "data-table", "empty-state", "error-state"],
+      "features": ["page-shell", "detail-pane", "status-chip", "data-table", "empty-state", "error-state", "filter-input", "chat-filter-namespace"],
       "screengrabCount": 0
     }
   ],
@@ -275,8 +275,8 @@
   ],
   "metadata": {
     "totalDocs": 30,
-    "totalScreengrabs": 48,
-    "featuresCovered": 82,
+    "totalScreengrabs": 42,
+    "featuresCovered": 95,
     "appSections": 16
   }
 }