We're live in beta — earn up to 12,000 credits by signing up today. Get started

FloConnector MCP
for AI agents

FloConnector's own connector. Enable it on a profile and Claude can write the skills your AI clients follow: describe a procedure in the conversation you are already having, and it lands in your workspace as a draft, with a version history. It acts as the person calling it, within their role.

Beta. Enable it today and every tool below is live. It is new enough that we are still testing it against real workspaces, so tell us if something looks off and we will fix it fast.

Built in-house. There is no third platform in the picture: this is FloConnector's own control plane, exposed as tools, so nothing leaves us and there is no credential to hold on your behalf.

Single connection. One FloConnector organisation per workspace. Connect another in a separate workspace.

Tools

Everything your AI can do in FloConnector

14 tools your agent gets the moment FloConnector is connected.

14 tools

  • skill_guide read

    Skill writing guide

    Read this BEFORE creating or updating a skill. What a skill is (SKILL.md body + references/, scripts/, assets/), the size limits, how flc_skill_create / flc_skill_update / flc_skill_publish fit together, and the draft rule.

  • collection_list read

    List collections

    The workspace's skill collections with skill counts by status and which profiles carry each. A collection is a folder of skills a profile exposes.

  • skill_list read

    List skills

    Skills in the workspace (or one collection) with status, tool name, current version and file paths. Drafts and archived skills included — unlike flc_about, which lists only published ones.

  • skill_get read

    Read skill

    A skill's body (or one of its files with `file`), at the current or a past `version`, with its metadata, file list and current version number — the number to pass as expected_version on flc_skill_update.

  • skill_create write

    Create skill

    Create a skill as a DRAFT (v1) in one call: name, description (when to use it), the markdown body (no frontmatter needed; up to 20480 bytes) and optional files in references/, scripts/ or assets/ (up to 25, 32768 bytes each, text only). Read flc_skill_guide first. Publish separately with flc_skill_publish.

  • skill_update write

    Update skill

    Revise a skill. ONE call is ONE revision, so send the body and every file change together. `files` is a DELTA — only the paths you are changing, `content: null` to delete. Always say what changed in `changelog`; pass `expected_version` from flc_skill_get so you never overwrite unseen work. Name/description changes alter the tool clients see. Rules and limits: flc_skill_guide.

  • skill_publish write

    Publish skill

    Publish a draft (it becomes a skill_… tool in every profile carrying its collection) or unpublish an active skill back to draft. Needs skills.publish — an admin by default.

  • skill_archive write

    Archive skill

    Archive a skill (reversible; it exposes nothing) or bring an archived one back as a draft. Permanent deletion is dashboard-only.

  • skill_history read

    Skill history

    Every revision of a skill, newest first: version, author, how it arrived (editor / MCP client / import / restore), the note, and the documents it touched.

  • skill_restore write

    Restore skill version

    Put a past version back — body and files together — as a NEW revision. Never rewinds; the history keeps everything.

  • skill_copy write

    Copy skill

    Duplicate a skill (body and files) into another collection as a draft. A taken slug gets a _copy suffix.

  • skill_export read

    Export skill

    The skill as an Agent Skills bundle: a directory listing with SKILL.md (spec frontmatter, version under metadata) and every file, ready to recreate and zip for a client that installs bundles.

  • collection_create write

    Create collection

    Create a skill collection. Its slug (from the name) becomes the tool prefix skill_<slug>_ for every skill in it and cannot change later. Needs skills.manage.

  • collection_update write

    Update collection

    Rename or re-describe a collection, change its emoji, or archive/restore it (status). The slug never changes. Needs skills.manage.

What it covers

FloConnector through FloConnector

FloConnector's own connector, and the only one that points back at us rather than out at a platform. Knowledge is where a workspace writes down how it actually works: how a lead is qualified, how a quote is worded, which checks a month-end close runs. Enabling this connector puts those documents within reach of the AI client already open on your desk, so the procedures can be written and revised in the conversation that prompted them instead of in a separate tab afterwards.

Every tool acts as the person whose AI client is calling it and checks the same permissions the dashboard checks, so enabling it grants nobody access they did not already have. New skills arrive as drafts and reach no AI client until someone publishes them.

Things people ask for

  • Write a skill for how we qualify an inbound lead
  • Add a pricing-objection section to the sales skill
  • Revise the onboarding skill and say what changed
  • Show me what changed in the month-end close skill this week
  • Publish the draft we wrote yesterday

Questions, answered

What is the FloConnector connector?

It is FloConnector's own control plane, exposed as tools. Enable it on a profile and Claude can list your skill collections, draft a new skill from the conversation you are already having, revise one, and publish it, without leaving the chat.

How do I turn it on?

Create a FloConnector workspace and enable FloConnector under Connectors: there is no account to sign in to and nothing to paste, because the connector is FloConnector itself. Add it to a profile, then start a new session in your AI client so it picks up the new tools.

Can it do things I have not allowed?

No. Every call runs as the signed-in person and checks the same permissions the dashboard checks, so enabling it grants nobody anything new. A tool whose permission someone lacks is not offered to them, and new skills arrive as drafts that reach no AI client until somebody publishes them.

Does it cost anything to use?

No. Calls to this connector are included: they appear in your activity log like any other call, and are never metered.

Does this work with ChatGPT as well as Claude?

Yes. The endpoint is a standard hosted MCP server, so it works in any client that supports MCP connectors, including Claude and ChatGPT.