SideButton Marketing Website Knowledge Pack Files

Prerequisites

• Authenticated session (portal login required)
• At least one agent online (check "N agents · N online" counter)
• Cookie consent dismissed
• Note: metrics are live — never assert exact CPU/MEM/LOAD values, only verify elements exist with numeric values

Phase 1: Page Load & Structure

#	Test	Method	Pass Criteria
1.1	Agents page loads	navigate to /portal/agents	H1 "Agent Fleet" visible
1.2	Agent count summary	snapshot	"N agents · N online" text below heading (2-part dot-separated); "· N error" appended only when error agents exist
1.3	Share button	snapshot	button "Share" visible in top-right area
1.4	Run Workflow button	snapshot	Orange button "Run Workflow" in top-right
1.5	Table headers	snapshot	Column headers: AGENT, HEALTH, ACTIVITY, SEEN
1.6	At least one agent card	snapshot	Agent card with screenshot/avatar, hostname, IP, health metrics, status badges, action buttons

Phase 2: Agent Card — Identity & Badges

#	Test	Method	Pass Criteria
2.1	Screenshot thumbnail	snapshot	button "Click to view full screenshot" with img "{name} screenshot" — OR placeholder icon if no screenshot
2.2	Agent avatar	snapshot	img "{agent-name}" icon next to hostname
2.3	Agent hostname	snapshot	Bold hostname text, e.g. "sidebutton-agent-1"
2.4	Agent IP	snapshot	Grey IP text next to hostname, e.g. "46.224.17.8"
2.5	Role badges	snapshot/screenshot	SE, QA, SD, PM colored pill badges visible on agent card
2.6	SHARED badge	snapshot/screenshot	Sky/blue "SHARED" badge (sky-400 color) on shared agents (e.g. venmate agents)
2.7	Copy agent name	click button "Copy agent name"	No error, clipboard action triggered
2.8	Copy IP	click button "Copy IP"	No error, clipboard action triggered

Phase 3: Agent Card — Health Metrics

#	Test	Method	Pass Criteria
3.1	CPU metric	snapshot	span "CPU usage" present with percentage value
3.2	Memory metric	snapshot	span "Memory: N/N MB" present (e.g. "Memory: 1361/3819 MB")
3.3	Load metric	snapshot	span "Load avg: N.NN" present with numeric load average
3.4	Disk metric	snapshot	span "Disk usage" present with percentage value
3.5	Swap metric	snapshot	span "Swap: N/N MB" present (e.g. "Swap: 460/4095 MB")
3.6	Software versions	snapshot/screenshot	Version badges visible on some agents: CLAUDE, NODE, NPM, SB with version numbers
3.7	Refresh button	click button "Refresh agent data" → snapshot	No error; metrics elements still present (values may change)

Phase 4: Agent Card — Status & Activity

#	Test	Method	Pass Criteria
4.1	Online status	snapshot/screenshot	Green dot + "Online" text on connected agents
4.2	Offline status	snapshot/screenshot	Grey dot + "Offline" text on disconnected agents
4.3	IDLE badge	snapshot/screenshot	Grey "IDLE" badge on online agents not running jobs
4.4	BUSY badge	snapshot/screenshot	Orange "BUSY" badge on agents running a job
4.5	Effort badge	snapshot/screenshot	Effort level badge (e.g. "MAX") shown on busy agents
4.6	Running job link	snapshot	Active agent shows link "Running · {workflow} — queued" pointing to /portal/jobs/{id}
4.7	Queued job labels	snapshot	"Queued · {workflow}" labels for pending jobs
4.8	Overflow indicator	snapshot	"+N more" text when multiple jobs queued
4.9	Last seen	snapshot	Relative time displayed, e.g. "Just now"
4.10	Disabled Run Job	snapshot	Offline agents have button "Run Job" with [disabled] attribute
4.11	Queue link	snapshot	link "Queue" or link "Queue N" present, links to /portal/queue?agent={id}
4.12	Queue badge count	snapshot/screenshot	Queue link shows count badge when items are queued (e.g. "Queue 5")

Phase 5: Run Workflow Modal (top-right button)

#	Test	Method	Pass Criteria
5.1	Open modal	click button "Run Workflow" → snapshot	Two-panel modal appears: workflow picker left, config right
5.2	Workflow groups	snapshot/screenshot left panel	Workflows grouped by role: OPS, SE, QA, SD, PM with section headers
5.3	Select OPS workflow	click "Ops: Pull All Repos" → snapshot	Right panel shows title, description, no required input fields
5.4	Select SE workflow	click "SE Agent: Implement Fix + PR" → screenshot	Right panel shows "Ticket Url" input (orange border, "Jira ticket URL" helper) + "Hint" textarea
5.5	Workflow description	screenshot right panel	Short description text below workflow title (e.g. "Read ticket, understand context...")
5.6	Hint textarea	snapshot	Label "Hint (optional)", helper "Optional extra instructions for the agent", placeholder "e.g. Focus on the login module, skip tests..."
5.7	Auto pick default	snapshot	"Auto pick" button with lightning icon, selected by default (orange)
5.8	All N agents button	snapshot	"All N agents" broadcast button available next to Auto pick (only when agents exist)
5.9	Agent selection grid	snapshot	Individual agent buttons with colored dots: green = online, orange = busy, grey = offline (only when agents exist)
5.10	Disabled offline agents	snapshot	Offline agent buttons are [disabled] in the agent grid
5.11	Select specific agent	click an online agent button → snapshot	Agent button becomes selected (highlighted)
5.12	Entry Path dropdown	snapshot	ENTRY PATH section with "Agent default" combobox/dropdown (only when agents exist)
5.13	Effort level toggles	snapshot/screenshot	EFFORT LEVEL section: Max, High, Medium (default, orange) toggle buttons
5.14	No agents role hint	snapshot (with 0 matching agents)	"No agents with {role} role" text next to Auto pick (role matches selected workflow)
5.15	Effort toggle interaction	click "Max" → screenshot, click "Medium" → screenshot	Selected button gets orange highlight, others are dark
5.16	Run disabled without required fields	select SE workflow, leave Ticket Url empty → snapshot	Run button greyed out / disabled
5.17	Close modal (X)	click X button → snapshot	Modal dismissed, agents page visible
5.18	Close modal (Escape)	press Escape → snapshot	Modal dismissed, agents page visible

Phase 6: Run Job (per-agent button)

#	Test	Method	Pass Criteria
6.1	Open modal	click button "Run Job" on an online agent → snapshot	Same modal opens
6.2	Agent pre-selected	snapshot/screenshot right panel	Only the target agent button shown and selected (no Auto pick, no All agents, no other agents)
6.3	Effort level visible	screenshot	EFFORT LEVEL section visible with Max / High / Medium
6.4	Entry Path visible	screenshot	ENTRY PATH dropdown visible with "Agent default"
6.5	Close modal	press Escape → snapshot	Modal dismissed
6.6	Disabled on offline	snapshot offline agent card	button "Run Job" has [disabled] attribute, cannot be clicked

Phase 7: Screenshot Viewer

#	Test	Method	Pass Criteria
7.1	Open viewer	click button "Click to view full screenshot" → screenshot	Full-screen dark overlay with large screenshot image
7.2	Screenshot image	snapshot	img "Agent screenshot" element present in overlay
7.3	Agent name label	screenshot	Agent hostname displayed below the screenshot
7.4	Close button	snapshot	button "Close" visible in top-right
7.5	Close via button	click button "Close" → snapshot	Overlay dismissed, agents page visible
7.6	Close via Escape	open viewer → press Escape → snapshot	Overlay dismissed
7.7	No screenshot agent	check agent without screenshot thumbnail	Placeholder icon (monitor) shown instead of thumbnail, no "Click to view full screenshot" button

Phase 8: Share Button

#	Test	Method	Pass Criteria
8.1	Share button visible	snapshot	button "Share" present in top-right area
8.2	Click share	click button "Share"	No error, link copied to clipboard (no modal opens)

Phase 9: Cross-Module Navigation

#	Test	Method	Pass Criteria
9.1	Queue link navigation	click link "Queue" on agent → verify URL	Navigates to /portal/queue?agent={id}
9.2	Running job link	click running job link → verify URL	Navigates to /portal/jobs/{id}
9.3	Sidebar navigation	click sidebar links	All portal nav links work (Chat, Jobs, Queue, Workflows, Settings)

Phase 10: Empty State (0 Agents)

#	Test	Method	Pass Criteria
10.1	Empty page heading	navigate to /portal/agents (0 agents) → snapshot	H1 "Agent Fleet" visible
10.2	Zero count	snapshot	"0 agents · 0 online" text below heading (2-part, no error part)
10.3	Empty state message	snapshot/screenshot	Robot icon + "No agents registered yet" heading + "Agents appear here after their first heartbeat" hint text
10.4	No table headers	snapshot	No AGENT/HEALTH/ACTIVITY/SEEN column headers rendered
10.5	Share button present	snapshot	button "Share" still available on empty page
10.6	Run Workflow available	snapshot	button "Run Workflow" still clickable on empty page
10.7	Modal with 0 agents	click "Run Workflow" → screenshot	Modal opens; only Auto pick shown, no agent grid, no "All N agents", no Entry Path
10.8	Role-specific hint in modal	select SE workflow → screenshot	"No agents with se role" text next to Auto pick
10.9	OPS role hint	select OPS workflow → screenshot	"No agents with ops role" text next to Auto pick
10.10	Effort level in empty modal	screenshot	Effort toggles (Max/High/Medium) still visible; Medium selected by default

Phase 11: Error Status & Version Drift

#	Test	Method	Pass Criteria
11.1	Error status badge	snapshot (agent with error)	Red dot + "Error" text on agent with status='error'
11.2	Error in count	snapshot (fleet with errors)	"· N error" appended to header count text
11.3	Error + IDLE	snapshot	Agent with error status shows "IDLE" badge alongside red "Error" dot
11.4	Share count badge	snapshot (owned shared agent)	Sky badge with share icon + count number (e.g. "2") on agents shared with other accounts
11.5	Version drift warning	snapshot/screenshot	Warning triangle icon next to outdated version badge when agent version < fleet max
11.6	All version badges	snapshot	CLAUDE CODE, NODE, NPM, SB version badges with version numbers

Phase 12: Edge Cases

#	Test	Method	Pass Criteria
12.1	Metrics refresh over time	snapshot → wait 30s → snapshot	CPU/MEM/LOAD values may change (live polling data)
12.2	Multiple agents	snapshot	Page renders all agent cards (scroll to verify all visible)
12.3	Mixed online/offline	snapshot	Online and offline agents render differently (disabled buttons, status indicators)
12.4	Agents without screenshots	snapshot	Agents without screenshots show placeholder icon, no "Click to view full screenshot" button
12.5	Agents without version info	snapshot	Some agents lack software version badges — no error, row simply absent
12.6	Waiting job status	snapshot (agent with waiting job)	Amber "Waiting" badge + "Waiting · {workflow}" link on agent with waiting job

Automation Tips

• Health metrics are live — don't assert exact percentages, just verify elements exist with numeric values
• The Run Workflow modal workflow list uses plain button elements, not links — click by ref
• After opening Run Workflow modal, the workflow buttons are anonymous button refs — map by position within the modal (OPS first, then SE, QA, SD, PM)
• The screenshot viewer overlay may be dismissed by snapshot — use screenshot tool to capture evidence of it
• "BUSY" badge only appears when agent is running a job — may not be testable on demand; verify on agents currently running jobs
• Offline agents have disabled Run Job buttons — test with actually offline agents
• DO NOT submit the Run button with real workflows in production testing — only verify button state (enabled/disabled)
• For the screenshot viewer, always use screenshot before snapshot since snapshot may dismiss the overlay
• Queue link count badge appears dynamically — only visible when agent has queued items
• "All N agents" count reflects currently online agents, not total agents
• Default effort level is Medium (orange highlight) — not High. Verify Medium is pre-selected on modal open
• "No agents with {role} role" is a dynamic hint that changes per workflow — test by selecting different workflow roles
• With 0 agents: modal hides agent grid, "All N agents", and Entry Path — only Auto pick + role hint + effort toggles shown
• Agent count format is conditional: "N agents · N online" baseline, "· N error" only appended when errors exist — don't assert 3 parts unconditionally
• SHARED badge is sky/blue color (sky-400), not green — verify badge color matches _skill.md
• Error status and version drift may not always be testable — only verify when agents in those states exist