| title | ERP•AI Skills |
|---|---|
| description | Open enterprise skill stack for business, implementation, research, and platform work. |
| audience | both |
| category | index |
ERP•AI Skills is an open library of operational playbooks for humans and AI agents. It covers department processes, implementation templates, investment research, platform operations, and founder work.
| Area | Purpose | Count |
|---|---|---|
departments/ |
Department and org-size specific business skills | 181 |
templates/ |
Cross-department implementation artifacts by org size | 12 |
investment-research/ |
Research craft plus sector-specific enterprise-tech coverage | 14 |
platforms/ |
Product-specific operating playbooks, currently build.host | 43 |
entrepreneurship/ |
Early-stage venture operating manual | 1 |
Total: 251 SKILL.md files.
skills/
├── departments/
│ ├── customer-support/
│ ├── finance-accounting/
│ ├── human-resources/
│ ├── information-technology/
│ ├── project-operations/
│ ├── sales-crm/
│ └── supply-chain/
├── templates/
├── investment-research/
├── platforms/
│ └── build-host/
└── entrepreneurship/
Canonical leaf paths:
departments/<department>/<org-size>/<skill>/SKILL.md
templates/<org-size>/<template>/SKILL.md
investment-research/core/<skill>/SKILL.md
investment-research/sectors/<sector>/SKILL.md
platforms/<platform>/<category>/<skill>/SKILL.md
entrepreneurship/<concept>/SKILL.md
The size tier refers to the whole organization's headcount, not the size of the department.
| Folder | Meaning |
|---|---|
01-org-under-100 |
Fewer than 100 people |
02-org-100-to-1k |
100 to 1,000 people |
03-org-1k-plus |
1,000+ people |
Use the tier that matches the organization today. A 5,000-person company with a 10-person HR team still uses 03-org-1k-plus.
Each department has an OVERVIEW.md that explains shared entities, integrations, metrics, and operating context. Read it before opening an individual skill.
| Department | Overview | Scope |
|---|---|---|
| Customer Support | OVERVIEW.md | Ticketing, SLA, escalation, knowledge base, omnichannel, field service, CSAT |
| Finance & Accounting | OVERVIEW.md | GL, AP, AR, fixed assets, budgeting, close, consolidation, tax |
| Human Resources | OVERVIEW.md | Payroll, benefits, recruitment, onboarding, offboarding, leave, performance, training |
| Information Technology | OVERVIEW.md | API design, integrations, data, security, observability, DR, testing, go-live |
| Project Operations | OVERVIEW.md | Planning, resourcing, time and expense, billing, portfolios, subcontractors |
| Sales & CRM | OVERVIEW.md | Leads, pipeline, CPQ, contracts, commissions, territory, customer 360, campaigns |
| Supply Chain | OVERVIEW.md | Procurement, inventory, warehouse, planning, logistics, quality, vendors, production |
departments/information-technology/role-overviews/ contains role references for solution architects, requirements analysts, data engineers, migration architects, compliance analysts, and QA leads. They are references, not installable skills.
Templates are cross-department artifacts scoped by org size:
fit-gap-matrixgo-live-checklistmigration-runbookrequirements-traceability
Example path: templates/03-org-1k-plus/go-live-checklist/SKILL.md.
investment-research/ combines research craft with enterprise-tech sector coverage.
core/contains research workflow skills such as filing analysis, earnings calls, valuation, competitive landscape, recommendations, and publishing.sectors/contains sector skills and ticker watchlists for HR tech, finance/accounting tech, sales/CRM tech, customer-support tech, supply-chain tech, project-ops tech, and platform IT.
Start at investment-research/OVERVIEW.md.
platforms/build-host/ contains agent playbooks for deploying, operating, designing, and scaffolding projects on build.host. Start at platforms/build-host/OVERVIEW.md.
Every SKILL.md uses YAML frontmatter:
---
name: accounts-receivable
description: "Use this skill when managing customer invoices, payment collection, cash application, aging, and dunning."
version: 1.0.0
agents:
- collections
- reconciliation
related:
- general-ledger
- period-close
metadata:
author: erphq
domain: erpai.studio
department: finance-accounting
size_tier: 03-org-1k-plus
type: skill
scope: internal
---Required fields:
namedescriptionversionmetadata.authormetadata.domainmetadata.typemetadata.scope
Optional downstream-facing fields:
agents: agent capability slugs that work on this skill.related: other skill slugs this skill commonly works with.
related slugs must resolve to another SKILL.md in this repo. Consumers that mirror only part of the repo should hide unresolved related slugs rather than render broken links.
erphq/lab-sites currently mirrors a curated subset of department skills under:
apps/web-lab/src/content/skills/<department>/<size-tier>/<slug>/SKILL.md
The current renderer keys those mirrored pages by bare skill slug. That works for the curated enterprise mirror, but the full repo must be keyed by canonical path because many slugs intentionally repeat across org-size tiers, such as:
departments/finance-accounting/01-org-under-100/accounts-payable/SKILL.md
departments/finance-accounting/02-org-100-to-1k/accounts-payable/SKILL.md
departments/finance-accounting/03-org-1k-plus/accounts-payable/SKILL.md
Full-repo consumers should treat canonical path as identity and name as the local skill slug.
Run:
node scripts/validate-repo.mjs
gitleaks detect --source . --config .gitleaks.toml --redact --verboseThe validator checks:
- all
SKILL.mdfiles have required frontmatter and an H1 - YAML plain scalars do not contain unquoted
:values agentsandrelatedare lists when presentrelatedreferences resolve somewhere in the repo- Markdown and HTML local links resolve
- current lab-sites enterprise department slugs stay unique
- legacy brand strings and live-looking example tokens do not reappear
CI runs the same repo validation on pull requests.
The enterprise tier (03-org-1k-plus) is populated across all departments. Information Technology now has full 01 / 02 / 03 coverage, Project Operations has full 01 / 02 / 03 coverage, and the current ERP.ai app registry resolves against the tier-specific Supply Chain leaves it references. Some non-enterprise folders still contain only .gitkeep until there is a concrete skill to write.
MIT. Use it, fork it, and build with it.