Claude Code
Work on tasks with unified Floor context from Claude Code on your computer.
Lit Factory MCP
Solutions · Connected tools
Keep context with the work, across the tools your team already uses.
Lit Factory connects operational tools through MCP so people and agents can work from shared, source-linked context without replacing the systems your team relies on.
The operating problem
A logo wall does not explain how a decision gets better.
The useful question is not how many tools appear in a catalog. It is which source holds the evidence for this role, what the agent can prepare from it, and which consequential action remains a human decision.
Connected evidence used across delivery decisions
LinearPriorities, tickets, ownership, and delivery state
GitHubImplementation, review, and delivery activity
SlackInterruptions, decisions, and handoffs
FigmaDesign context and open product decisions
SentryOperational signals that change delivery risk
PostHogProduct usage, adoption, and behavioral signals
HubSpotContacts, companies, deals, tasks, and pipeline state
Lit Factory connects operational tools through MCP. Available context and actions depend on the connections and permissions configured for your workspace.
First-class AI coding agent support
Bring the operating context into the coding agent.
Claude Code, Codex, and Cursor can read Floor state and contribute through the Lit Factory MCP server. Tickets, decisions, ownership, and review context travel with the task instead of being reconstructed in a separate coding session.
Codex
Work on tasks with unified Floor context from Codex on your computer.
Lit Factory MCP
Cursor
Work on tasks with unified Floor context inside Cursor on your computer.
Lit Factory MCP
First-class support does not mean blanket authority. Available context and actions depend on the configured client, workspace connections, source permissions, and the human autonomy settings attached to the work.
Connection evidence and limits
Connected toolContext used hereWhat Lit Factory preparesPublic boundary
Linear
Context used here
Priorities, tickets, ownership, and delivery state
What Lit Factory prepares
Planning, implementation, risk, and review briefs
Public boundary
Specific writeback depends on the configured MCP service and granted permissions.
Jira
Context used here
Planned work, ownership, and delivery state
What Lit Factory prepares
Ticket-to-delivery context where Jira is the work source
Public boundary
Shown as an alternative work source; universal Jira actions are not claimed.
GitHub
Context used here
Implementation, checks, review, and delivery activity
What Lit Factory prepares
Actionable review, change context, and release evidence
Public boundary
Repository access and any prepared follow-through remain permission-bound.
Slack
Context used here
Interruptions, decisions, clarifications, and handoffs
What Lit Factory prepares
Decision recaps, response drafts, and owner handoffs
Public boundary
A prepared message is not represented as sent before human approval.
Figma
Context used here
Design context and open product or experience decisions
What Lit Factory prepares
Planning context and consequential design-delta briefs
Public boundary
Direct design-file editing is not claimed on this site.
Sentry
Context used here
Operational signals that can change delivery posture
What Lit Factory prepares
Operational-risk and release-posture briefs
Public boundary
Correlation is evidence, not proof; automated remediation is not claimed.
PostHog
Context used here
Product usage, adoption, and behavioral signals
What Lit Factory prepares
Planning evidence, product-signal briefs, and operating-review context
Public boundary
Available evidence depends on the connected PostHog project and granted permissions; product changes are not implied.
HubSpot
Context used here
Contacts, companies, deals, tasks, and pipeline state
What Lit Factory prepares
Customer context, planning evidence, follow-up, and pipeline actions
Public boundary
CRM reads and writes depend on the connected account and user-granted MCP permissions; consequential changes remain human-controlled.
Lit Factory connects operational tools through MCP. This matrix describes the bounded evidence patterns presented on this site, not every object or action an MCP service might expose. Availability, freshness, and actions depend on the configured service and workspace permissions.
One agentic workflow
Carry a delivery decision across systems without losing its evidence.
A tool connection contributes bounded context or an allowed action. The Floor keeps the cross-tool operating story visible while each source remains the system of record.
Trigger01
Connected work needs a cross-tool decision.
A ticket, review, interruption, design change, or operational signal marks the work for attention.
Grounded context02
Only relevant sources are gathered.
The agent follows configured MCP connections and workspace permissions to assemble the evidence needed for this decision.
Prepared action03
A bounded next move is proposed.
The agent summarizes the evidence, names uncertainty, and prepares a source-specific update, handoff, or follow-up.
Human decision04
A person authorizes consequence.
The owner approves, changes, or rejects the proposed move. Lit Factory records the decision with its operating context.
After trust is established
Automate the repeatable follow-through.
After a team trusts a specific pattern, automate routine synchronization, routing, and reminders within the permissions already granted. New tools and higher-consequence actions remain explicit choices.
Human in Control
The decision stays human.
MCP provides the connection surface; it does not grant blanket authority. Workspace connections, source permissions, and human decisions define what context an agent can use and what actions it can take.
Intended outcome: connected evidence becomes a traceable operating decision, not another disconnected notification.