191 pre-validated AACP v1.1 encoding rules for common enterprise workflows.
MIT Licensed. Maintained by the AACP spec author. The canonical implementation of the community encoding registry called for in the AACP specification.
A rule is a pre-built, validated AACP v1.1 packet for a common business workflow step. Instead of calling an LLM to encode a natural language instruction, your agent loads the matching rule from this library and gets the packet at zero cost.
{
"id": "hr-merge-calculate-payroll",
"name": "Merge and calculate payroll run",
"description": "Merge employee salary data with cost centre budgets and calculate full payroll",
"task": "MERGE",
"dom": "HR",
"packet": "MERGE|HR|return:HR-Agent|p:1|aacp:1.1|rules:payroll_v2|validate:budget_cc",
"tags": ["payroll", "merge", "calculate", "paye", "pension"],
"version": "1.1",
"source": "community",
"validated": true
}| Domain | Rules | Workflows covered |
|---|---|---|
| HR | 38 | Payroll, onboarding, absence, performance, headcount, contractors, offboarding |
| FIN | 35 | Invoicing, budget, month-end close, expenses, payment runs, FX, VAT, reporting |
| IT | 30 | Provisioning, access, incidents, licences, vulnerabilities, devices, change |
| SALES | 25 | Lead qualification, pipeline, forecasting, renewals, commissions, competitive |
| CS | 23 | Ticket resolution, sentiment, churn, NPS, SLA, goodwill |
| LEGAL | 22 | Contracts, NDA, GDPR, regulatory, litigation, policy, IP |
| MKT | 18 | Campaigns, lead quality, content, attribution, A/B testing |
Every rule is validated against the AACP v1.1 schema.
"validated": true on every rule in this library.
import json
from pathlib import Path
# Load a specific rule
rule = json.loads(Path("rules/HR/hr-merge-calculate-payroll.json").read_text())
packet = rule["packet"]
# Send packet to your HR agentfrom aacp import RuleRegistry
registry = RuleRegistry.from_community()
rule = registry.find("payroll merge calculate")
print(rule.packet)
# MERGE|HR|return:HR-Agent|p:1|aacp:1.1|rules:payroll_v2|validate:budget_ccimport json
index = json.loads(Path("index.json").read_text())
rules_by_domain = {}
for rule in index["rules"]:
rules_by_domain.setdefault(rule["dom"], []).append(rule)Your agents are writing coordination messages in natural language. Every one of those messages is an API call and tokens cost money. This library gives you 191 pre-built, validated packet templates for the most common enterprise workflow steps -- payroll, invoicing, IT provisioning, contract review, ticket handling, sales qualification.
Drop the matching rule into your agent's coordination message instead of natural language. Zero API cost for encoding. Same packet every time. Machine-readable. Auditable.
For anything not covered, AACP's fallback encoder calls an LLM once, logs the result, and serves it from cache on every subsequent call.
rules/
HR/ 38 rules — payroll, onboarding, absence, performance
FIN/ 35 rules — invoicing, budgets, month-end, expenses
IT/ 30 rules — provisioning, access, incidents, licences
SALES/ 25 rules — leads, pipeline, forecasting, renewals
CS/ 23 rules — tickets, sentiment, churn, NPS, SLA
LEGAL/ 22 rules — contracts, GDPR, regulatory, IP
MKT/ 18 rules — campaigns, attribution, content
schema/
rule.schema.json JSON schema for rule validation
index.json Full library as single file (191 rules)
CONTRIBUTING.md How to submit new rules
Every rule in this library meets these requirements:
- Validated against AACP v1.1 schema (
"validated": true) - TASK is one of the 12 valid values
- DOM is one of the 7 valid values
return:field presentaacp:1.1field present- Packet under 200 characters
- No empty positional slots
- Description matches what the packet does
Rules failing validation are not merged. CI runs validation on every PR.
See CONTRIBUTING.md for the rule submission process. Rule submissions are welcome for any domain and workflow type.
- Spec and SDK: https://aacp.dev
- Python SDK: https://github.com/MackayAndrew/aacp (
pip install aacp) - TypeScript SDK: https://github.com/MackayAndrew/aacp-ts (
npm install aacp-ts) - IETF Draft: https://datatracker.ietf.org/doc/draft-mackay-aacp/
MIT — free to use, copy, modify, and distribute.