# Access Pledge

> Decide your event's accessibility provisions, publish them on one page, and answer every request. Enter who funds it,
> whether the venue is signed, the needs already known, who owns accessibility and when the set must be published, and
> get the provision set decided - depth, captioning and interpretation, the request channel and the disability data it
> creates, economic access and the slot itself - then the one public accessibility page, the request form, and an answer
> to each request. Built on the SkillSafe skill @samber/event-accessibility-inclusion, from
> samber/dev-event-organizer-skills (MIT), by Samuel Berthe.
> It decides the bar, not delivery; it is not the code-of-conduct response team; it is not legal advice - what a
> jurisdiction requires is a question for counsel. It never asks for or repeats a diagnosis.

https://access-pledge.skillsafe.ai/

## What is free
- The funding pole's consequences (a community budget moves the depth, captioning and economic menus, never the request channel) and the venue's (unsigned: the criteria sheet for event-venue-sourcing first; signed: what the building lacks is deleted and stated).
- The four menus re-ranked for the interview with every move named and every ruled-out rung struck: provision depth, captioning and interpretation, the request channel and the data it creates, and economic access after the slot check.
- The provision set - physical, sensory, dietary, economic, digital - each with its status, depth and sizing, owner and funding, known needs first; the people attendees turn to; the WCAG 2.2 AA digital-layer checklist; the data-handling posture; each pasted request classified (covered, covered with lead time, not covered - decline with reason, needs owner decision) with its lead time against the event date; the hand-offs and the stated gaps; a verdict (answer the requests, fix first, decide before the venue is signed, ready to publish).
- Exports: the venue criteria sheet (.md), the provision set (.csv), the draft page skeleton (.md), everything (.json).

## The metered lanes
Two tasks on gpt-terra, metered per run, routed by `task`. Input string fields: `task` (plan or page), `event`, `needs`, `requests` (may be empty) and `facts`.
- `plan` replies with `summary`, `verdict`, `menus`, `slot_note`, `provisions`, `people`, `digital`, `handoffs`, `data`, `gaps` and `caveats`; the page holds the picks, the provision rows, the checklist and the hand-offs to its own.
- `page` replies with `title`, `intro`, `sections` (one per area, covered and not covered), `request_channel`, `day_of_contact`, `last_updated`, `form`, `answers` (one per pasted request, never repeating a diagnosis), `disclaimer` and `note`.

## Links
- API tutorial: https://access-pledge.skillsafe.ai/api.html
- Source skill: https://skillsafe.ai/skill/@samber/event-accessibility-inclusion