You finished the Foundation module with three documents: a competency-framework.md, an inclusive-jd.md template, and a sensitive-comms.md for your team. This toolkit makes them infrastructure. A job description you wrote once fills one role; a competency framework your team installs — that Claude reads on every task — shapes every hire, every review, and every promotion conversation, forever.
Everything here is copy-able. Replace the Mizan placeholders (the GCC bookkeeping SaaS whose own HR team runs through the module) with your own, save the files in the folder your team works in, and open it in Claude Desktop. From then on, every people task in that folder starts from your real role definitions, your real seniority standards, and your real communication norms — instead of each hire re-deciding them.
The whole pack is four files and a handful of prompts. That’s the point — a foundation small enough to actually get installed beats a beautiful one that lives in a shared drive nobody opens. Keep the sensitive data out of it: this folder holds your people infrastructure, never a single compensation figure or medical record.
The HR CLAUDE.md
This is the standing context Claude reads automatically when it opens your project folder, so nobody has to re-explain the team setup every chat. Keep it short and current — it points at the three foundation docs rather than repeating them. The fairness gate at the bottom is not a disclaimer; it’s the discipline that makes it safe to use Claude on people decisions at all.
# CLAUDE.md — [Company] people & HR
## The company
[Company] is [one-line — what we do]. We're [N] people, [UAE-based],
[Series A]. The HR team is [2 people]: [HR manager] and [ops generalist].
We support [N] staff across [functions].
## The three source-of-truth files — read these first
- competency-framework.md → WHAT each level means (Specialist / Senior / Lead
and above, mapped to observable behaviors per
competency dimension)
- inclusive-jd.md → HOW we write job descriptions (structure, must-haves
vs. preferences, the compensation block, what to
leave out)
- sensitive-comms.md → HOW we communicate difficult decisions (structure,
register, what not to open with)
Always read the relevant ones before drafting. If a JD, review, or message
conflicts with these files, the files win — flag the conflict, don't quietly
rewrite the standard.
## How we work with you
- You draft; a person decides. Surface your reasoning (why a behavior maps to
a level, why a requirement is genuine vs. a proxy) so a human can check it.
State nothing as fact you can't show from the files or what I've told you.
- Apply competency-framework.md exactly — never re-infer a level by feel.
- Bilingual work: ask which language leads before drafting.
## The fairness gate (never skip)
- Any output that names or describes a specific person — a review, a
performance message, a hiring assessment — is reviewed by a human before
any action. Claude drafts, never decides about people. A draft that maps a
named employee to a level or recommends an outcome goes to the HR manager
for review before it goes anywhere else.
## What's sensitive — Desktop file pane only
- Sensitive: compensation data, salary bands, offer letters, medical or leave
records, performance improvement plans, personal contact detail, visa or
immigration status.
- How we handle it: Claude reads the local copy via the Desktop file pane;
each read is approved in the "Ask permissions" prompt. Never paste sensitive
data into the chat window.
- Safe in this folder: the framework, template, and comms docs. Company
headcount and structure. Anonymized role-level aggregates (e.g., "3 Leads,
5 Seniors" — no names).On Desktop you create and edit this in the file pane — no commands. Anthropic’s docs call the “Ask permissions” mode the one recommended for new users; leave it on, and the first time Claude reads this file you’ll approve it in the prompt.
The competency-framework.md template
The shared definition of what each level means at your company — kept in one place so every JD, every review cycle, and every promotion conversation starts from the same bar. The “how you’d see it” column is the discipline: a behavior described in the abstract gets re-interpreted under pressure; one grounded in an observable moment holds.
# competency-framework.md — [Company]
## Seniority levels
Specialist → Senior → Lead (→ Principal / Head, role-dependent)
Default to Specialist for new-to-function hires; Senior for 3–5 yrs of
domain depth; Lead for cross-team scope or people management.
## Dimensions × levels
### 1. Role execution (doing the core job well)
| Level | Observable behavior | How you'd see it |
|------------|-----------------------------------------------------------|----------------------------------------------------|
| Specialist | Completes defined tasks reliably; asks when unclear | Delivers a clean client onboarding; flags blockers |
| Senior | Owns a domain end-to-end; spots edge cases unasked | Catches a data-import gap before the client does |
| Lead | Sets the standard for the function; fixes broken systems | Writes the onboarding runbook the team follows |
### 2. Client / stakeholder relationship
| Level | Observable behavior | How you'd see it |
|------------|-----------------------------------------------------------|-----------------------------------------------------|
| Specialist | Responds reliably; escalates with context | Sends a status update before being chased |
| Senior | Manages expectations proactively; handles objections | Reframes a scope change without losing trust |
| Lead | Builds strategic relationships; seen as a trusted advisor | Gets pulled into the client's planning conversations |
### 3. Communication
| Level | Observable behavior | How you'd see it |
|------------|-----------------------------------------------------------|-----------------------------------------------------|
| Specialist | Clear written and verbal; suited to audience | Summarizes a long thread in two sentences |
| Senior | Shapes others' understanding; adapts register | Explains a technical error to a non-technical CFO |
| Lead | Sets communication norms for the team; writes for record | Drafts the post-mortem format everyone else uses |
### 4. Judgment & problem-solving
| Level | Observable behavior | How you'd see it |
|------------|-----------------------------------------------------------|-----------------------------------------------------|
| Specialist | Applies known frameworks; escalates novel problems | Uses the categorization rules doc before asking |
| Senior | Diagnoses root cause; weighs trade-offs independently | Spots that a recurring issue is a process gap |
| Lead | Frames ambiguous problems; decides under uncertainty | Calls a process change before being asked to |
### 5. Leadership & developing others
| Level | Observable behavior | How you'd see it |
|------------|-----------------------------------------------------------|-----------------------------------------------------|
| Specialist | Contributes to team knowledge; asks good questions | Writes up what they learned for the shared notes |
| Senior | Onboards peers; identifies gaps in the team's capability | Runs a session on the new filing process |
| Lead | Grows people; owns the team's output, not just their own | Blocks time to review a Specialist's client work |
## Gray areas, decided once
- [the ambiguous case] → [level], NOT [the other]. Reason: [why].
- ...
## What this framework is not
It is not a scorecard that auto-decides promotions. It is the shared vocabulary
a human uses to have a fair, grounded conversation. The HR manager reviews any
output that maps a named person to a level before that output goes anywhere.The inclusive-jd.md template
The structure that keeps job descriptions honest — must-haves only in the requirements block, preferences clearly separated, and compensation visible. The “what not to do” notes are as important as the sections; they’re the decisions that get re-litigated under hiring pressure if they aren’t written down.
# inclusive-jd.md — [Company] · [Role title]
## Role outcome
[One sentence: what does success look like in 12 months?]
e.g. "You own the onboarding experience for 30+ new Mizan clients per quarter,
from contract-signed to first reconciliation submitted."
WHAT NOT TO DO: Don't open with a paragraph about us ("Mizan is a fast-growing
SaaS..."). Lead with what the person achieves, not what we are.
## What you'll do
- [Outcome-oriented bullet — what you produce, not just what you do]
- [3–5 bullets maximum — if you need more, the scope is two roles]
- ...
WHAT NOT TO DO: Avoid process bullets ("attend weekly standups", "use Jira").
Process is how, not what. If a bullet can apply to any job at any company, cut it.
## Genuine requirements (must-haves only)
- [Concrete skill or credential that is genuinely necessary, not a proxy]
- [Keep this list short — 3–4 items. Every extra line narrows the pool without
improving it. "Bachelor's degree" is almost never a genuine requirement.]
- ...
WHAT NOT TO DO: Don't list years of experience as a requirement unless the role
is legally scoped to it. "5+ years" as a bar excludes capable people and adds
nothing the competency framework doesn't already describe. Use the framework to
describe the level; let the framework do that work.
## Things that help but aren't barriers (we'd love, not required)
- [Genuine preference — useful but not a dealbreaker]
- [Keep to 2–3 items; more blurs the must-have line]
- ...
WHAT NOT TO DO: Don't list preferences that are actually requirements under a
softer label. If missing it would rule someone out, move it to must-haves.
## Compensation and location
[AED X,XXX – X,XXX / month] · [Location: Dubai, hybrid / remote]
Visa [provided / not provided].
WHAT NOT TO DO: Don't omit the range. A missing range signals either that we
don't know what the role is worth, or that we're negotiating against the
candidate from the start. Publish the real band.The sensitive-comms.md template
The fill-in framework for messages that carry real weight — restructuring, performance conversations, role changes, difficult feedback. The structure is deliberate: the decision first, the uncertainty named honestly, and the next step clear. Everything else — the rationale, the context, the “we’ve thought hard about this” — comes after, or not at all.
# sensitive-comms.md — [Company] · [message type]
## 1. The decision (lead with it — one sentence)
[What is actually happening, plainly stated.]
e.g. "We are restructuring the implementation team, and your role is being
made redundant, effective [date]."
WHAT NOT TO DO: Don't open with "I wanted to be transparent with you" or "After
a lot of careful consideration." These are a tell that the writer is about to
bury the lede. The person reading already knows something is coming; starting
with context before the decision prolongs the anxiety and erodes trust.
## 2. What this means for you
- [The concrete practical consequence — money, timeline, next step]
- [A second consequence if relevant]
- [Keep to what is known and confirmed; don't speculate]
## 3. What we're still figuring out
We don't yet know: [name the open question honestly].
We'll have an answer by [date / "the end of this week"].
WHAT NOT TO DO: Don't fill this section with vague reassurances if there's real
uncertainty. "We're committed to supporting you" without a concrete next step
reads as a platitude. Name what's unresolved; set a date for resolution.
## 4. What hasn't changed
[What the person can rely on — their relationship with their manager, their
remaining project ownership, the team's respect for their work.]
This section matters most in restructuring and role-change messages, where people
need an anchor. It's optional in pure-feedback contexts.
## 5. The next step
[One clear action, with a name and a date: "We'll meet on [date] to walk through
the transition plan. I'll send the calendar invite today."]
NOTES ON REGISTER
- Write the way the person's manager actually speaks — not legal-register, not
HR-register. If the manager says "I" not "the company," write "I."
- Read the draft aloud. If it sounds like a press release, rewrite it.
- The HR manager reviews every draft before it reaches the recipient.
Claude drafts; a person delivers.The saved-prompt library (Foundation)
The recurring foundation jobs, as prompts you reuse. Paste them in the chat with the right files open; on the Terminal & Automation track these become slash commands so they’re one keystroke, but you need none of that to use them today.
Write the competency framework from a job description
"Read this job description [open or paste it]. Reverse-engineer the observable
behaviors implied by each responsibility — what a Specialist, Senior, and Lead
would produce differently on this role. Then draft a competency-framework.md
section using our five dimensions: role execution, client relationship,
communication, judgment, leadership. Flag any dimension the JD doesn't give
you enough signal on, and ask me to fill it in."
Write an inclusive JD from the competency framework
"Read competency-framework.md and inclusive-jd.md. I want to hire a [level]
[role title]. Using the competency-framework level definitions as the source
of truth for the bar, draft a job description in the inclusive-jd.md structure:
role outcome, what you'll do, genuine requirements only (no years of experience
unless legally required), preferences clearly separated, and a compensation
band placeholder I'll fill in. Flag any requirement you're unsure is genuine
vs. a proxy."
Draft a sensitive message for [situation]
"Read sensitive-comms.md. I need to write a [restructuring / performance /
role-change / difficult-feedback] message to [role, not name]. The facts:
[what happened / what is changing / what is decided]. Draft a message in the
sensitive-comms.md structure — decision first, what it means, what's uncertain,
what hasn't changed, next step. Flag anything I've said that's vague or that
you'd need to know more about before the message holds."
Pressure-test a JD for exclusionary language
"Read this job description [open the file]. Check it against inclusive-jd.md:
are the requirements genuine must-haves or are some of them proxies? Is the
experience bar stated in years instead of competencies? Are there requirements
in the preferences section that would actually disqualify someone? List what
to cut, move, or rewrite — don't just describe the problem."
Refresh the foundation
"Our team has changed: [new role / seniority restructure / team reorg]. Read
competency-framework.md, inclusive-jd.md, and sensitive-comms.md and tell me
which definitions are now stale, what to update, and propose a versioned diff.
Don't rewrite silently — show me the before/after."Installing it (Desktop, no setup)
- Save the four files —
CLAUDE.md,competency-framework.md,inclusive-jd.md,sensitive-comms.md— in the folder your HR team works in. - Open that folder in Claude Desktop and approve the read in the “Ask permissions” prompt.
- Start any task by opening the relevant docs; the prompts above do the rest.
That’s the whole install. The foundation you built in the module is now shared infrastructure — and Module 2 of the track extends this same pack with the hiring pipeline and review-cycle templates that build on it. If you’re standing this up across a team rather than for yourself, read the operating guide for the data, permissions, and rollout layer that belongs underneath.