mbfinotti Business Skills - Setup Guide¶
Source: mbfinotti/sales-skills, advertising-skills, revops-skills, partnerships-skills (0-1 stars each, MIT LICENSE; last pushed Sep 14, 2026) Skill family: mbfinotti business-skills family - 4 repos, 108 skills (sales 27, advertising 31, revops 20, partnerships 30) Combined Installs: ~637,148 across 108 skills in 4 repos (Oct 8, 2026 snapshot) Category: Business Operations / Sales, RevOps & Advertising Quality Tier: Beta - high-quality operator content, MIT, interview-based design, single-author with modest stars; verified Oct 8, 2026
mbfinotti is a four-repo, 108-skill family for running the revenue side of a business: sales-skills (27) for running a sales org and closing deals, advertising-skills (31) for spending paid media budget on purpose, revops-skills (20) for the systems behind the number, and partnerships-skills (30) for channel, affiliate, referral, and creator operations. The author (Maya-Beth Finotti) built the collections for colleagues at Nativa Labs, and the design is consistent across all four: every skill is interview-driven - it asks multiple-choice clarifying questions before acting - and every skill ends in a decision or a ship-able artifact (a scorecard, a routing spec, a metric tree, a rate card), never a generic explainer. Coverage spans market sizing, ICP, quota and comp design, MEDDPICC and negotiation in sales; channel selection, budget policy, audience design, creative development, and measurement across search, paid social, video, native, and AI-answer surfaces in advertising; funnel design, lead scoring and routing, pipeline hygiene, forecasting, health scoring, and CRM data governance in RevOps; and partner economics, co-selling, affiliate program mechanics, referral incentives, and influencer deals in partnerships. The evidence standard is real: alliance-prioritization cites Kale, Dyer & Singh (2002), finding 63% vs 50% alliance success across 1,572 alliances, and diagnosis skills such as ad-account-diagnostic return a prioritised verdict with evidence and confidence.
Installation¶
# Full family (recommended - install all four repos, not single skills)
npx skills add mbfinotti/sales-skills
npx skills add mbfinotti/advertising-skills
npx skills add mbfinotti/revops-skills
npx skills add mbfinotti/partnerships-skills
# Or pick a single skill
npx skills add mbfinotti/sales-skills --skill sales-pipeline-coverage-modeling
The READMEs ask for a full-repo install: the skills are atomic by design and reference each other freely, so a single-skill install leaves sibling references and routed handoffs pointing at skills that are not present. Host-specific install paths for Claude.ai, Claude Code, Codex, Cursor, and Gemini CLI are documented in each README.
Roster - mbfinotti/sales-skills (27 skills, ~208,882 installs)¶
The sell-side surface, from market sizing to negotiation:
| Skill | Installs | What It Does |
|---|---|---|
| sales-pipeline-coverage-modeling | 7,877 | Models how much pipeline the team needs relative to quota: coverage ratios from win rates, segment targets, and pipeline-gap math with in-quarter timing |
| sales-icp-definition | 7,888 | Defines the ICP with weighted scoring criteria, explicit disqualifiers, and a refresh cadence, from closed-won data or founder discovery |
| sales-market-sizing | 7,864 | Estimates TAM, SAM, and SOM by triangulating top-down, bottom-up, and value-theory methods, capped by sales-capacity limits |
| sales-account-tiering | 7,852 | Sets tier cutoffs above a fit score, with accounts-per-rep caps, a coverage model per tier, and a recalibration cadence |
| sales-career | 7,849 | Coaches a candidate through breaking into sales, SDR-to-AE promotion, interview prep, and evaluating an offer against dated benchmarks |
| sales-kickoff | 7,817 | Router: sends any broad sales request to the one skill that fits and bootstraps the project's shared sales-context file |
| sales-hiring | 7,851 | Builds the employer-side hiring loop: outcome scorecard, structured interview bank, scored mock-call work sample, 30-60-90 ramp plan |
| sales-objection-handling | 7,690 | Diagnoses what an objection really means - writes rebuttals a rep can say out loud, or calls the deal dead when the constraint is real |
| sales-radar | 7,838 | Assembles a dated, verified watch list of sales podcasts, newsletters, communities, and benchmark reports matched to your role |
| sales-account-segmentation | 7,654 | Builds the weighted account fit score from firmographic, technographic, and intent signals, calibrated on closed-won deals |
| sales-meeting-recap | 7,670 | Converts raw call notes into a recap email where every next step carries an owner, a date, and a deliverable |
| cold-email-deliverability | 7,759 | Audits sending setup and draft mechanics for inbox placement: SPF/DKIM/DMARC alignment, sender reputation, body hygiene, regional compliance |
Roster - mbfinotti/advertising-skills (31 skills, ~184,239 installs)¶
Paid media work across search, paid social, video, native, and AI-answer surfaces:
| Skill | Installs | What It Does |
|---|---|---|
| ad-creative-brief | 5,934 | Turns a campaign goal and audience insight into a brief a designer or creator can execute, with specs and a revision loop |
| ad-account-diagnostic | 5,914 | Diagnoses the root cause of an underperforming ad account and returns a prioritised verdict with evidence and confidence |
| ad-spend-allocation | 5,910 | Splits a fixed budget across campaigns, platforms, funnel stages, and audiences based on expected marginal return |
| retargeting-funnel | 6,048 | Designs a staged retargeting sequence with recency windows, behavioural depth tiers, a message ladder, and per-stage frequency caps |
| advertising-career | 6,165 | Plans a paid media career: the junior-to-lead ladder, interview formats, an NDA-safe portfolio, and pay conversations |
| ad-conversion-tracking | 5,929 | Verifies conversion events fire once and deduplicate before launch, then turns the evidence into a GO or NO-GO decision |
| paid-media-scaling | 6,080 | Decides when a proven campaign has earned a budget increase, how large each step should be, and what triggers a rollback |
| ad-budget-pacing | 5,914 | Tracks spend against budget and flags under- or over-pacing early: pacing ratio, projected spend, corrective daily spend |
| ad-swipe-file | 5,908 | Builds a queryable library of competitors' running ads by format, hook, offer, and funnel stage, then ranks test hypotheses |
| ad-audience-targeting | 5,895 | Turns an ICP and buying signals into layered audience tiers sized against platform floors and exclusion rules |
| ad-platform-selection | 5,906 | Decides which paid channel families fit the business economics, audience, funnel stage, and budget before campaigns get built |
| ad-creative-test-plan | 5,924 | Designs a pre-launch creative test with a falsifiable hypothesis, per-cell budgets, required sample, and kill or scale rules |
Roster - mbfinotti/revops-skills (20 skills, ~125,281 installs)¶
The systems behind the number, from funnel design to CRM governance:
| Skill | Installs | What It Does |
|---|---|---|
| sales-forecast-diagnostic | 6,297 | Diagnoses why a forecast misses, separating data-quality problems from rep behaviour from genuine demand, and recommends targeted fixes |
| pipeline-stage-definition-audit | 6,322 | Audits stage definitions against buyer-verifiable milestones and flags exit criteria built on rep activity instead |
| revenue-funnel | 6,325 | Designs a revenue funnel model from scratch: stage set, unit of analysis, conversion assumptions, ownership handoffs |
| revops-radar | 6,297 | Assembles a time-budgeted watch list of RevOps podcasts, newsletters, communities, events, and people, each verified active |
| revops-hiring | 6,292 | Produces a complete RevOps hiring packet: outcome scorecard, interview stage map, work-sample rubric, and a 30-60-90 ramp |
| revenue-kpi-framework | 6,313 | Designs the org-wide KPI tree: metric math from board to IC, branch ownership, and guardrail counter-metrics per owned number |
| revops-kickoff | 6,315 | Router: sends any RevOps task to the right skill and bootstraps a versioned project context so later sessions start warm |
| crm-data-governance | 6,372 | Produces CRM field-level governance: field dictionary, per-field ownership, write-precedence rules, freshness SLAs, enforcement |
| revops-career | 6,320 | Builds a RevOps career plan: ladder placement by scope, competency gap roadmap, evidence ledger, and a compensation ask |
| sales-pipeline-hygiene | 6,324 | Runs a checklist audit over a live pipeline snapshot and returns an exception list, a disposition per deal, and a pass threshold |
| revenue-leakage | 6,314 | Traces where deals silently exit one funnel and sizes the loss in recoverable dollars, separating leaks from healthy disqualification |
| revenue-data-governance-strategy | 6,261 | Sets org-wide source-of-truth policy per object class, an identity-resolution spine, and a data-contract register |
Roster - mbfinotti/partnerships-skills (30 skills, ~118,746 installs)¶
Channel, alliance, affiliate, referral, and creator operations:
| Skill | Installs | What It Does |
|---|---|---|
| alliance-prioritization | 4,073 | Ranks named candidate alliances by expected value, effort, and risk into a shortlist and a go/no-go recommendation memo |
| partner-ecosystem-expansion | 3,959 | Sequences which partner categories to launch next into a staged roadmap with readiness gates, capacity limits, and kill criteria |
| influencer-campaign-brief | 3,953 | Writes the creative brief for a signed creator: deliverable specs, key messages, guardrails, disclosure, bounded review rounds |
| partnerships-kickoff | 3,947 | Router: sends a partnerships task to exactly one sibling skill and bootstraps the shared project context file |
| co-selling-strategy | 3,945 | Defines how direct and partner sellers share deals: registration policy, deal credit splits, rules of engagement, comp neutrality |
| influencer-discovery-brief | 3,944 | Defines weighted sourcing criteria and knockout screens, then produces a ranked creator shortlist with evidence per score |
| affiliate-payout-audit | 4,102 | Audits a commission payout run before disbursement and returns findings by severity with recommended holds |
| partner-economics | 3,932 | Models one partner's P&L - margin, cost-to-serve, partner CAC, ramp, payback - to decide sign, scale, renegotiate, or exit |
| partner-tiering | 3,929 | Designs the tier ladder inside an existing program: qualification criteria, benefit bundles, promotion and demotion rules |
| partner-channel-conflict | 3,929 | Writes the channel conflict rules for contested deals: account segmentation, carve-outs, tie-breaks, escalation ladders |
| partnerships-hiring | 3,928 | Plans partnerships hiring: job posting and scorecard, interview loop, sourcing, and compensation stance |
| partner-channel-program | 3,927 | Designs a partner program from scratch: readiness gate, partner value proposition, motions, tiers, benefits, economics envelope |
Why This Matters for Hermes Agents¶
These are operator-level business workflows - pipeline coverage modeling, ICP definition, ad account diagnosis, partnership and affiliate operations - which is exactly the kind of work business teams bring to AI agents once the agent can hold context and produce artifacts. Each skill runs as a short interview and returns something ship-able (a scorecard, routing spec, concession plan, or rate card), so a Hermes agent can run a sales, RevOps, advertising, or partnerships engagement end to end without improvising its own structure. Because sibling skills read the same shared context file, a session that starts at a kickoff router can hand off to pipeline, comp, or negotiation skills with the context intact.
Verification¶
# Confirm the family is installed (all four repos should appear)
npx skills list | grep mbfinotti
# Smoke-test the interview flow with your agent host
# Prompt: "Run sales-pipeline-coverage-modeling for a 6-rep team carrying a $4M quota"
# Expected: clarifying questions first, then a coverage model with segment-level gap math.
Security¶
skills.sh per-skill verdicts (sampled 3 skills, verified Oct 8, 2026). All sampled verdicts were Pass across scanners; verdicts vary per skill, so check a skill's security page on skills.sh before production use:
| Skill | Gen Agent Trust Hub | Socket | Snyk |
|---|---|---|---|
| sales-pipeline-coverage-modeling | Pass | Pass | Pass |
| ad-creative-brief | Pass | Pass | Pass |
| alliance-prioritization | Pass | Pass | Pass |
Integration with CorpusIQ¶
CorpusIQ is the connector layer that lets AI assistants read the systems a business already runs on - accounting, payments, CRM, email, files and analytics - and answer with the source named. It is read-only by design: it retrieves data and never writes to, or acts in, a connected system, which is what makes it safe to point at live accounts.
This page is part of the CorpusIQ directory, the reference we maintain for the MCP server and agent-skill ecosystem. Run alongside CorpusIQ, a third-party server like this one can be combined with a company's own business data in a single answer.
Limitations¶
- Verified Oct 8, 2026: four MIT repos, created Sep 5, 2026, first seen on skills.sh Sep 13, 2026, and last pushed Sep 14, 2026; 0-1 stars and 0-1 forks each.
- Single-author family (mbfinotti, built per the READMEs for colleagues at Nativa Labs). Modest community traction: the case for installing rests on content quality and the interview-based design, not stars.
- Dormant since Sep 14, 2026, so platform-adjacent details in the skills (ad platform behavior, CRM fields) should be sanity-checked against the current tools.
- Install counts are a single Oct 8, 2026 skills.sh snapshot; the ~637,148 combined figure (108 skills) moves with the index, so treat it as approximate.
- Roster tables show 12 representative skills per repo; the full repos ship 27 / 31 / 20 / 30 skills.
- No live install test was performed as part of this guide; design details come from the repo READMEs and the Oct 8, 2026 data snapshot.
Related¶
- HubSpot Agent CLI Skills - CRM Operations Setup
- Digital Marketing Pro - AI Marketing OS Setup
- Apify Growth Skills - Lead Gen and Brand Monitoring
- Skills Catalog
- Skills Marketplace
Pro Tips¶
- Install whole repos, not single skills. Each collection is designed to cross-reference; a single-skill install breaks routed handoffs to sibling skills.
- Start at the kickoff router. sales-kickoff, advertising-kickoff, revops-kickoff, and partnerships-kickoff route the request and bootstrap the shared context file that makes later sessions start warm.
- Answer the clarifying questions. The multiple-choice interview is what keeps outputs specific to your situation; telling the agent to skip it degrades the artifact toward generic advice.
- Feed real data first. Point the agent at your CRM export, ad platform report, or partner tracker before running a skill; the models and scorecards are only as good as the inputs.
- Sample before you standardize. Run one skill per repo against live data, review the artifact quality, then decide which of the four repos to roll out to the team.
← Skills Catalog | Skills Marketplace → Powered by CorpusIQ