# Mentor Roster

> Enough mentors for every hackathon track, reachable when a team is stuck. Enter the hackathon brief's challenge
> tracks and sponsor integration, the hacking window and rooms, and the mentors you have with what each actually knows,
> and get coverage depth set per track rather than a headcount, the recruitment channels your roster deadline allows,
> every mentor tagged in the brief's vocabulary with the mentor-judge conflict handled, the gaps by track and hour -
> then the request mechanic teams will read and the mentor briefing with the help-versus-build boundary. Built on the
> SkillSafe skill @samber/hackathon-mentoring, from samber/dev-event-organizer-skills (MIT), by Samuel Berthe.
> A mentor is not a volunteer, and not a judge. It never invents a mentor-to-participant ratio, a staffing formula or
> a response-time target, hands shift sizing to the volunteer plan as text, and never infers anyone's identity or
> seniority from a name - mentored before is asked.

https://mentor-roster.skillsafe.ai/

## What is free
- Coverage depth per challenge track, set against the tracks rather than a headcount: the rung (rostered mentor shifts, an expertise-matched roster, on-call from organizers, no mentor programme) and the commitment (guaranteed or available), with a sole-substrate or required sponsor API raising that track's bar to a mentor guaranteed on the floor whenever it is worked.
- Every mentor tagged in the brief's own vocabulary, the sponsor technology by product name, built-with versus read-the-docs depth, and every unmatched tag listed - never dropped; a mentor whose tags match no track is flagged.
- The coverage hour by hour across the hacking window, every gap by track and hour, and the overnight hours called out with the decision to publish (staffed, thinly staffed, unstaffed).
- The mentor-judge conflict stated from one side: the contact log whenever a mentor also judges; the recusal rule and its mechanics stay with the judging plan.
- The recruitment, request-mechanic and briefing-depth menus re-ranked for the interview, every move named and every deleted rung struck; the channels the roster deadline allows with their lead times; the help-versus-build boundary and the sponsor engagement rule.
- Rotation by expertise, interruption discipline, and the expertise constraint handed to the volunteers' shift plan as text - never a computed rota. A verdict: map tracks first, fix coverage, fix first, ready to publish the mechanic.
- Exports: the coverage grid (.csv, track by hour), the mentor tags (.csv), every mentor's slots (.ics), a mechanic draft (.md), everything (.json).

## The metered lanes
Two tasks on gpt-terra, metered per run, routed by `task`. Input string fields: `task` (plan or brief), `event`, `tracks`, `mentors`, `notes` (may be empty) and `facts`.
- `plan` replies with `summary`, `verdict`, `coverage`, `tags`, `gaps`, `conflict`, `recruitment`, `rotation`, `handoff_to_shifts` and `caveats`; the page holds every rung, tag, gap and channel to its own.
- `brief` replies with `mechanic` (title, how to ask, where it is published, the hours, what to expect) and `briefing` (the boundary, the conflict note, interruption, escalation, the rotation note, the mentors per track), and a `note`.

## Links
- API tutorial: https://mentor-roster.skillsafe.ai/api.html
- Source skill: https://skillsafe.ai/skill/@samber/hackathon-mentoring
- The same hackathon: https://team-forge.skillsafe.ai/ (team formation), https://jury-room.skillsafe.ai/ (judging) and https://prize-shape.skillsafe.ai/ (prizes)