nha-in/uhi-integrators-assistant

v0.1.2MIT

Run /uhi-start after installing: it says what this plugin does, offers prompts to start from, and asks what you are building. Agent skills for UHI integration from the ABDM Developer Portal, compiled from the Catalogue. One skill per service: uhi-consultation (Physical Consultation, as EUA or HSPA), uhi-ambulance (Ambulance Booking, as EUA or HSPA), uhi-pmjay-hem (PM-JAY hospital search, as EUA), uhi-blood-bank (Blood Bank Discovery, as EUA or HSPA), uhi-jan-aushadhi (Jan Aushadhi Kendras and medicines, as EUA) and uhi-notto (transplant hospitals, as EUA). Each carries its work as sections that load on demand: scaffold builds the service journey by journey against the sandbox, design holds what the screens around the calls must do, integrate holds the hosts, headers, signing and the context block, debug walks a failed call or a missing callback to a fix proved by the original step succeeding, and test runs the service's test cases and the go-live steps. Every step names the operation it came from. Three commands open the work from the terminal: /uhi-preflight checks what must be true before the first call, /uhi-prove-signing signs one PM-JAY HEM search and proves the header before a flow is built on it, and /uhi-decode-response turns an HTTP status, an ACK or NACK, an error object or a silence into its most likely cause and the next check. Two agents plan above the skills: uhi-integration-agent takes a goal such as making a clinic bookable, picks the role and the service skills and sequences them by what each requires and produces, and uhi-call-debugger takes one failing call or one missing callback to a fix and verifies it by the original step succeeding. An EUA completes Milestone 2 on HIE-CM first, which the abdm-integrators-assistant plugin carries. Every skill declares its domain, what it requires and produces, and that it never orchestrates.

GitHub Open Website

UHI Integrators Assistant

Three layers, each answering one question.

Layer	Question	Lives in
Agent	What am I trying to accomplish, and what should I do next?	`agents/`
Skill	What is the correct way to perform this task?	`skills/`
Tool	What concrete operation can I execute?	HTTP, signing, the network registry, the codebase, the sandbox

Commands under commands/ are thin entry points into a skill or an agent, never a fourth layer.

First run

Run /uhi-start after installing. It says what the skills, agents and commands do, offers prompts to start from, and asks what you are building. scripts/build-skills.mjs writes it from the skills manifest and the frontmatter beside it, so edit those rather than commands/uhi-start.md. Each skill's SKILL.md carries its own Try asking section from the same source.

Agents

Two, and only two.

uhi-integration-agent takes a goal such as "make our clinic's doctors bookable", picks the role and the service skills it needs, orders them by their requires and produces, and holds each step to its exit condition.
uhi-call-debugger takes one failing call or one missing callback to a fix, verified by the original step succeeding. It reads each service's references/debug.md and carries no causes of its own.

Commands

/uhi-preflight [service] checks what must be true before the first call: subscriber ID, key pair, a reachable callback URL, fresh IDs, and for an EUA, Milestone 2 on HIE-CM.
/uhi-prove-signing signs one PM-JAY HEM search and proves the header against the sandbox Gateway before a flow is built on it.
/uhi-decode-response turns an HTTP status, an ACK or NACK, an error object or a silence into its most likely cause and the next check.

Skills

One folder per service: uhi-consultation, uhi-ambulance, uhi-pmjay-hem, uhi-blood-bank, uhi-jan-aushadhi and uhi-notto. A service is not one task. Its SKILL.md is a router and its capabilities sit under references/: scaffold builds the journeys one at a time, design holds what the screens must do, integrate holds the hosts, headers, signing and the context block, debug walks a failed call or a missing callback to a fix, and test runs the service's test cases and the go-live steps.

Each skill installs and runs alone. It repeats what it needs of signing, the registry lookup and the context block rather than depending on another skill.

Every skill declares its role in frontmatter: type: skill, its domain, the agent_consumers that may dispatch it, what it requires and produces, and can_orchestrate: false. A skill is a known procedure from a known input to a known output. It never plans.

The skill folders are generated from the UHI specification, the journey files and skills-src/. Edit those, never skills/.

Before any UHI service

An EUA completes Milestone 2 on HIE-CM before it can be onboarded onto any UHI service. The abdm-integrators-assistant plugin carries that work, in its abdm-m2 skill.

The decision rule

Create an agent only when the work has this shape: a goal, an unknown next action, inspect, choose, act, interpret, choose again. A known input through a known procedure to a known output is a skill.

Do not create an agent per service. No consultation agent, no blood bank agent. A user with a goal should face one agent and its relevant skills, not a choice between six.