Leadership · MCP tools
The Leadership Rhythm Kit — a mirror, not a generator.
Most leadership tooling promises to tell you what to say to someone. These three do something less comfortable and more useful: they tell you what you have been avoiding. Where your 1-on-1s never go. How the sentence you were about to send is going to land. What your read of a colleague says about your own behaviour.
They run inside whatever assistant you already use — Claude, ChatGPT, Cursor — through the A Founder's Friend MCP server. Every judgement is made strictly inside the frameworks published on this site, and every answer is about you.
Three tools, one story
design_one_on_one
The rhythm designer
Nine boxes, one session. It picks the one you keep skipping.
A 1-on-1 is three topics — how the work is going, how the person is doing, where they are heading — crossed with three perspectives: what I see, what you see, what we owe each other. Nine boxes. Most leads live in two of them and call it a cadence. The tool picks the box this session should sit in, tells you why, names the boxes that have gone cold, and hands you the questions that open that box rather than the ones you already ask.
Free to run. Coverage over time and the ledger are a supporter perk.
{"name":"design_one_on_one","arguments":{"subject_label":"the designer","situation":"Shipping fine, but quiet in group settings for a month.","covered_elsewhere":["daily standup","monthly roadmap review"]}}evaluate_feedback
The feedback evaluator
Read it back to yourself before the other person has to.
Paste the thing you were about to say. It comes back placed on Radical Candor — care personally, challenge directly — with the Nonviolent Communication parts you left out, a prediction of how it will actually land on the reception scale (Destroy, Deny, Defend, Dismiss, or the three responses you were hoping for), a score, and one rewrite in your own register. Sometimes the verdict is that this was never feedback: it was a coaching question wearing feedback's clothes.
Free to run. Quality trends over time are a supporter perk.
{"name":"evaluate_feedback","arguments":{"draft":"You keep dropping the ball on deadlines and everyone has noticed.","subject_label":"Person A","relationship":"report"}}read_the_room
Coach the coach
The hardest read in leadership is the one that includes you.
Describe a situation — no names — and it places the person on the awareness grid: oblivious, aware, or developing, working in isolation or inside a team. It tells you whether this is a coaching moment or a feedback moment, reads the situation against the 100% life capacity rule and the three-strikes trust formula, and then turns exactly one question back on your own behaviour. Every time. That last part is the point of the tool.
Free to run. Placement history is a supporter perk.
{"name":"read_the_room","arguments":{"situation":"Believes they are the strongest performer; the team quietly routes around them.","subject_label":"Person B","works_in":"team"}}leadership_ledger
The ledger
A mirror, not a generator.
Over a quarter, the kit accumulates the shape of your practice: which of the nine boxes you actually cover, which you have never once opened, how your feedback distributes across the Candor quadrants, whether your drafts are getting better or just faster. Nothing about any colleague — only what you did.
Supporters only.
{"name":"leadership_ledger","arguments":{}}This is the most sensitive data in a workplace
A team lead's private notes about a named colleague's performance, wellbeing and self-awareness are exactly the kind of record employment law is careful about. So three rules are enforced in code, not in a promise:
No names. Ever.
Every tool takes a label you choose — "Person A", "the designer", "my tech lead". Input that looks like a name or an email address is rejected, by the tool and again by the database. The mapping between label and person stays in your head, or in your employer's system. Never here.
No content at rest.
The feedback draft, the situation you describe, and the rewrite you get back are processed for that one answer and then gone. What can be stored is the shape of your own practice: which box, which quadrant, a score, a date.
Your data, on a clock.
Records are scoped to you and no one else — there is no team view, no manager-of-manager view, no sharing. They are deleted automatically after twelve months by default, three or six if you prefer, and immediately whenever you ask.
Access and erasure are free and need no email to anyone: export_my_leadership_data returns everything held about you as JSON, and erase_my_leadership_data with confirm: "ERASE" is a hard delete. Full detail on the privacy page.
If you lead at a larger company
Your employer, not this site, is the controller of performance data about your team — and it may have its own rules about tools like this. What to tell your HR partner or DPO, in one paragraph:
It is a coaching aid for my own practice. No colleague is identified: only a pseudonym I choose, and names and email addresses are rejected outright. Nothing I write about another person is retained — only which topic I covered, which feedback quadrant I landed in, a score and a date, tied to my own account. Records delete automatically after twelve months, and I can export or erase them myself at any time. It is not an HR record and nothing in it feeds a performance process.
What this deliberately is not
- —Team-wide dashboards
- —Manager-of-manager visibility
- —Sharing a ledger with anyone, including your employer
- —Employer or HR accounts
- —Uploads of HR exports or performance files
The moment a leadership tool can be read by someone above the lead, it stops being a mirror and becomes surveillance. That line is the product.
When you need names, not labels
This kit is a mirror on your own practice, so it never holds a colleague's name and never keeps what you wrote about them. If what you actually need is named, ongoing feedback and coaching across a team — with consent, retention and a legal basis your employer can stand behind — that is a different product, and two built for it are worth your time: Teamspective for continuous feedback and engagement, and Taito for structured coaching and development. Both are the right answer where privacy and employment law stop this tooling, by design.
Connect it
The endpoint, the client configuration, and the full tool reference live on the MCP page. In short:
https://kcrucyjpzabirewfvynq.supabase.co/functions/v1/mcpThe frameworks behind the judgements are all readable here: the 1-on-1 framework, the feedback framework, leadership development and thoughts on performance.