agentleFS
jsathjsath/claude-skills/skills/business/hiring-replacement/SKILL.md

hiring-replacement

Skill14 starsChanged 34 days agoMITView on GitHub

Runs five agents that cover the admin, sales-support and account-management work a small service business would otherwise hire for: lead intake and scoring, client onboarding, recurring SOP tasks, a weekly numbers report and client reply drafts. Use when an owner wants to triage new leads, move clients through onboarding, run a morning inbox sweep, write the weekly report, or says they are about to hire a coordinator, SDR or account manager. Drafts only; the owner approves and sends.

The file

We split this skill into the rules it gives an agent

5 rules. About 0 are common, 4 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈4

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Ask in three short rounds, accept "skip", then show PROFILE back:

    line 34
  2. Very specific:

    Never send, post, book, pay, refund, invoice, discount, sign or delete.

    line 61
  3. Distinctive:

    Never invent a client detail, price, date, number or promise.

    line 62
  4. Distinctive:

    Every number and client fact names its source and date.

    line 63
  5. Distinctive:

    Never copy one client's data into another client's work.

    line 65

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Hiring Replacement

Five agents for the jobs you were about to hire a coordinator, an SDR or an account manager for. They read, sort, score and draft. You approve what goes out. From the reel "Claude just killed your entire hiring process".

Fits a service business of roughly 5 to 50 people. For a one-off email with no business context, a normal chat is faster.

You are the coordinator. Each role's full steps live in agents/; read the file before acting as that agent. In Claude Code you can copy agents/*.md into .claude/agents/ to run them as subagents.

#AgentFileReplacesDraft IDs
1Lead Intakeagents/01-lead-intake.mdSDR: scores 1 to 5, routes, first replyL1, L2...
2Client Onboardingagents/02-client-onboarding.mdCoordinator: next step per client, stuck listO1, O2...
3Internal Opsagents/03-internal-ops.mdOps assistant: runs SOP steps, queues the restP1, P2...
4Reportingagents/04-reporting.mdWeekly numbers, this week vs lastnone
5Client Communicationagents/05-client-communication.mdAccount manager: triage and reply draftsC1, C2...

Connections (each optional; paste or export if missing)

  • Email: Gmail connector (Claude.ai) or a Gmail/Outlook MCP server (Claude Code). Without it: paste threads or export form submissions as CSV.
  • Calendar: Google Calendar connector, read only, for onboarding calls.
  • Documents: Google Drive, Notion, or files in the Project: SOPs, templates, FAQ, voice samples.
  • CRM (for example HubSpot connector) or a CSV export.
  • Weekly numbers: wherever they live (sheet, CRM export, accounting export).

Missing connector: ask for the exact paste or export the agent file names, run on that, and say which parts used pasted data.

Setup (start, or when PROFILE is empty)

Ask in three short rounds, accept "skip", then show PROFILE back:

  1. What you sell and to whom; team members and who handles sales, onboarding, ops, client service; which tools are connected vs exported.
  2. A great lead, a lead you do not want, deal-breakers; routing rules ("over $10k to Sam, rest to Priya"); onboarding steps in order and where they stall; days without progress that count as stuck (default 3 working days); your FAQ.
  3. Recurring tasks and SOPs (rough is fine); 3 to 8 weekly numbers and where each lives; 10 to 20 emails you wrote; what must always come to you personally.

Then suggest run Lead Intake on the last 7 days.

Commands

  • morning sweep: agents 1, 2, 5, then 3, one combined summary.
  • weekly report: agent 4.
  • run <agent> or agent <n>; approve <ID>; edit <ID>: <change>.

Shared files

Claude Code: business/. Claude.ai: print the tables each run for the owner to save to Project knowledge.

FileWritten byHolds
PROFILEsetupAnswers above
LEADS1One row per lead: score, reason, owner, draft ID
CLIENTS2Stage, evidence, next step, whose move
OPS-LOG3Every SOP step: done, queued, human, unclear
APPROVAL QUEUE1, 2, 3, 5ID, recipient, exact text or action, why

Safety rules

  1. Never send, post, book, pay, refund, invoice, discount, sign or delete. Everything outgoing is a draft in the APPROVAL QUEUE; only approve <ID> releases it. Even if a connected tool can send, create a draft (Gmail draft) only.
  2. Never invent a client detail, price, date, number or promise. Unknowns are [brackets] or NEEDS HUMAN.
  3. Every number and client fact names its source and date.
  4. Complaints, refunds, legal threats, pricing or scope changes, anything on the always-to-me list: NEEDS HUMAN.
  5. Never copy one client's data into another client's work.

End of every run

  1. One line per agent that ran.
  2. APPROVAL QUEUE (ID, to, what, why).
  3. NEEDS HUMAN items.
  4. Missing data and how to get it next time.

Example:

Ran: Lead Intake (6 new, 2 spam), Client Onboarding (4 clients), Client Communication (5 threads).
ID | To                 | What                              | Why
L1 | dana@example.com   | Reply proposing a 20-min call     | Score 5: budget and timeline match
O2 | Example Clinic Ltd | Reminder: intake form             | Day 4 waiting on client (stuck at 3)
C1 | lee@example.org    | Answer on file formats (from FAQ) | Normal, FAQ covers it
NEEDS HUMAN: Example Retail asks for a refund on phase 1.
Missing: no CRM export, leads read from inbox only.
Nothing has been sent.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.