Skill packs › Skill

Daily Sales HabitIN EVERY PACK

The Daily Sales Habit agent — the capstone. Runs on a schedule (that's the point), reads YOUR funnel and YOUR weekly MAA, and comes back each morning with ONE staged selling action - the email drafted, the post written, the follow-up list ready. Honor requested draft delivery; complete routine sending within recorded recipient/channel authority. Nothing invented; consequential decisions or missing authority need advance review.

Skill file sales-every-day.md · last updated Oct 7, 2026

Start with one result. Read this guide or find it in the skill-pack directory. Give your AI app the guide and the source files it can read, then ask for one draft you can check. A ZIP holds instructions; it does not install a plugin or grant access. For reusable setup, follow the current installation guide and verify a fresh run. This one ships in every pack.

The rule this agent enforces: you sell every day. Not "launch twice a year." Not "post when inspired." One deliberate selling action, every working day, aimed at the weakest stage of YOUR funnel. Most owners don't skip this because they disagree — they skip it because deciding what today's action is costs more willpower than doing it. So the agent prepares the finished asset, honors an explicit draft request, and completes routine delivery within recorded authority. You keep consequential decisions and control of the scope.

Read these first, every run: your brand-voice document, your ideal-client file, and your links file (real domains, offer pages, socials) — plus your latest weekly MAA (see the weekly-brand-maa skill). Everything below is grounded in those. If a file is missing, say which one and use the rest.

The funnel this agent watches

Lead magnet → Low-ticket offer (the small obvious next step, typically $7–197) → Core offer → Follow-up → Recurring/retainer. Revenue = traffic × conversion, stage by stage. Your weekly MAA names your weakest stage; this agent turns that diagnosis into daily motion. The most common gaps we see: a lead magnet and a core offer with nothing low-ticket in between, or no opt-in page at all. Local-business variant of the same funnel: free estimate → small paid diagnostic → the full job → review + referral ask → maintenance plan.

Why this agent is SCHEDULED (the whole point)

Every other skill waits for you to remember it. This one doesn't. Tell Claude once — "Create a scheduled task: every weekday at 7am, run my sales-every-day skill and leave today's action pack in my Outputs folder" — and from then on it behaves like a tiny sales department that reports for work before you're awake. That's the difference between having agents (you drive every run) and employing them (they run in loops and come back to you).

Hard rules

Selective review within recorded authority. Honor an explicit draft-only request. Routine factual follow-ups and approved publication may ship without per-item human review. New offers, pricing, commercial promises or missing recipient/channel authority require the decision owner's advance review. This skill does not authorize an unsolicited campaign or invent consent. Verify every sent or published result.
Nothing invented. Offers, prices, testimonials, links — only from your files or pages actually fetched and verified live. No fabricated urgency, no fake scarcity, no made-up bonuses.
Your voice. Brand-voice file governs. Style gate as a mechanical LAST pass: no "not just X, but Y", no em-dash soup, no AI-tell words, quotes verbatim.
One action per day. Never a to-do list of five things. The agent's discipline protects yours.
Log every run to Outputs/sales-every-day/log.md (date · funnel stage targeted · action staged · what you did with it). The agent reads the log first — it never repeats yesterday's action, and it notices when you haven't sent anything in 3 days (and says so, kindly, once).

Inputs

Latest weekly MAA — weakest stage + this week's focus (run the weekly-brand-maa skill if you don't have one).
Your knowledge base: brand-voice, ideal-client, links file, offers + prices if listed.
Your list size / email platform if connected (optional), your site (optional), your socials (optional).
The log from previous runs.

The daily loop (steps)

Read state. Log first. Then MAA + knowledge base. Identify: weakest funnel stage, this week's focus, what was staged yesterday and whether it went out.
Pick TODAY's action from the stage playbook below — the single highest-leverage move you can complete as a draft right now. Weekly rhythm guide: Mon plan + pipeline sweep · Tue content-to-offer · Wed list email · Thu follow-up day · Fri proof + MAA handoff. The stage playbook overrides the rhythm when a stage is BROKEN (no opt-in page beats everything).
Build the asset, completely. Not an outline — the finished draft, grounded in your files, links verified live.
Deliver in the requested mode. Honor explicit draft-only requests. For routine sending or publication with recorded recipient/channel authority and required QA, deliver and verify the result. A connector login or schedule alone grants no such authority; otherwise stage a finished draft and report its exact location.
Report in 5 lines: today's action · why this stage · verified delivery or draft location · any consequential decision or missing authority · yesterday's action status (sent/not).
Fridays only — pre-fill your weekly review. Three answers drafted from this week's log: what went out, what's next, where you had trouble — so the weekly MAA takes one minute, not twenty.

Stage playbooks (what "today's action" looks like)

No opt-in / lead-magnet page (fix FIRST — nothing enters an empty funnel):

Day 1: full opt-in page draft — headline, 3 bullets, form block, your lead magnet's promise in your voice.
Day 2: the lead magnet itself outlined from your best existing content (a video or post you already made).
Day 3: delivery email + welcome email drafts. Day 4: a social post driving to it. Day 5: MAA proof check.

Low-ticket gap (lead magnet + core offer but nothing in between — the #1 gap):

Day 1: low-ticket concept memo — 3 candidates carved FROM your core offer (a module, a workshop, a template pack, a paid diagnostic), each with price point and one-line promise; recommend one.
Day 2: its sales page draft. Day 3: the "P.S. upgrade" email to lead-magnet takers. Day 4: social proof post. Day 5: pipeline math — how many low-ticket buyers = one core-offer client.

No core-offer sales page: page draft from your testimonials + offer facts, then traffic moves.

Funnel complete → daily selling rhythm (rotate, never the same move twice in a row): one value email with a P.S. offer mention · one follow-up sweep (everyone who raised a hand in 30 days, with suggested one-line personal replies — sell without selling) · one same-day offer when the moment fits · one proof post (client win, verbatim quote) · one open-loop content post → lead magnet. For local businesses add: one review ask to a just-served customer · one quote-chase call list with talking points.

Output

Outputs/sales-every-day/<date>/ with the day's asset file(s) + report.md (the 5-line report), and the appended log line. Report the verified sent/published/draft status according to the requested mode and recorded authority.

Definition of done

One finished asset exists for today — routine authorized delivery is verified, or a requested draft/consequential decision is ready to act on.
It targets your CURRENT weakest stage, not a generic tip.
Zero invented facts/links/prices; style gate run LAST, mechanically.
The log line exists; the agent knows what tomorrow must not repeat.

Pairs with

→ weekly-brand-maa (the diagnosis this acts on) → measurement-analytics (real numbers instead of snapshots) → content-agent / content-factory (your raw material) → dollar-a-day-strategist (amplify the winners) → recursive-self-improvement-qa (grade the week, sharpen next week)

Built by Dennis Yu (Local Service Spotlight). Scheduled by design: a sales department that reports for work before you wake up, completes routine work within recorded authority, and reserves advance review for consequential decisions or missing authority.

Connecting to a CRM / marketing platform — and the one hard limit

Work at whatever tier the client has granted, and say which tier you used in every report.

Tier 0 — nothing connected. Read their files and public pages, produce finished drafts. Full value on day one.
Tier 1 — email platform. Honor requested draft-only delivery. Routine sending requires recorded recipient/channel authority; platform access alone does not grant it.
Tier 2 — their website. Use the requested delivery status within recorded client editing/publication authority. A WordPress application password alone does not grant publication authority.
Tier 3 — CRM. See below.
Tier 4 — ads + analytics. Read-and-recommend only. Never change a budget, never spend.

HighLevel / GoHighLevel (and white-labels of it): connect through HighLevel's official MCP server at https://services.leadconnectorhq.com/mcp/, authenticated with a Private Integration Token the client creates at Settings → Private Integrations in their own location. Never ask for, or accept, an account password — a PIT is scoped and revocable.

What the connector gives you (36 tools): contacts (create/update/upsert/tag/search), conversations (search, read, send), opportunities and pipelines, payments (orders, transactions), calendars, forms, custom fields, social posts, blog posts, and email templates (create + fetch).

What it does NOT give you — state this plainly rather than working around it: you cannot create or edit a funnel/landing page through HighLevel's API. Funnels are absent from the MCP entirely, and the REST Funnels API is read-only (list funnels, list pages, count pages). No create, no update, in any version. So: write the page onto the client's own website as a draft (better for their SEO anyway, and platform-neutral), and hand them paste-ready copy mapped to their page template. Never drive the CRM's UI by browser automation to fake a page build.

Send safety: conversations_send-a-new-message sends a real message to a real person and social-media-posting_create-post publishes. Honor an explicitly configured draft-only lane. Routine delivery may proceed within recorded recipient/channel authority after required QA; schedules and connector access grant no missing authority, and new offers or consequential decisions need advance review. emails_create-template creates a template and does not send a campaign.

Other skills: ai-search-visibility · boil-the-ocean · client-access-checklist · client-relationship-cadence · content-agent · content-factory

The full run order is on the skill-pack directory. Every skill here is one task from the Task Library — the library is the catalogue of what can be done; a pack is the subset you install; an agent is who runs it.

Where this sits in the system

Context is the information the worker needs. A skill describes how to do a task. A pack groups skills. Work happens when an authorized agent runs a job with the agreed inputs, checks the result, and saves the work record. Installing a skill does not start an agent or create a schedule.

Context — the verified facts, goals, evidence, decisions, and working state kept in user-owned Markdown, Obsidian, Drive, or Git so any authorized model can pick up the same work.
Skill — the written method for a task; use the linked Task Library (our directory of tasks and recipes) for its current tasks and state.
PackYOU ARE HERE — a selection of skills; follow the current installation guide for setup and access.
Agent — the AI worker carrying out an authorized task with the required access and tools.
Job — an authorized execution with a trigger, inputs, checks, and a saved result; recurring work also needs an actual configured schedule.
Proof — a written meta article (a record of one task run) for every execution, including partial or failed runs; public release follows recorded authority, and verified lessons improve the recipe.

The map: The System · every asset: Asset Tracker · next door: Asset Tracker.