Context bank

The course facts the team needs at hand, kept where an agent can read them

The context bank is the standing knowledge behind the tracker: what the project is and what has been decided, what the course is, when everything is due, what each deliverable must contain and how it is scored, who is on the team and who owns which account, and the conventions the professor grades on. It exists so that nobody (person or agent) has to re-derive these facts from the syllabus every time, and so that a fresh Claude Code session in this repo can read CLAUDE.md, then these pages, and be useful within a minute.

Page	What it holds	Source of truth
The project	What we are studying, attributes and levels, research questions, market, the decision log, the direction so far	The proposal draft and the team’s decisions
Course	What EMSE 6035 is, grading weights, toolchain	Syllabus and course site
Deadlines	The full semester calendar, every dated item	schedule.csv on the course site, mirrored in data/deadlines.csv
Project requirements	The seven deliverables, their sections and rubric points	The project pages and rubric CSVs
Team	Roster, roles, account ownership, contribution log	The team
Conventions	How this repo and site are maintained; the R and Quarto habits the course enforces	The professor’s notes plus team decisions

What does not belong here

This site is public because GitHub Pages needs a public repo. So: no grades or feedback, no email addresses, netIDs or phone numbers, no credentials of any kind (Supabase keys and shinyapps tokens live in a .Renviron on one teammate’s machine and nowhere else), and nothing about anyone’s work outside the course. If a fact fails that test it stays in a private place and this site links to nothing.