Files a repo can add

Any repo you own controls how it appears on the site through its own files. Only a README.md is required. Everything is read at build time over raw.githubusercontent.com (one git-trees call per repo powers the docs site and the blog autoscan), so there is no webhook or registration: add a file, push, and the next deploy picks it up. There are no per-project files in the website repo anymore; it all lives in the repo.

README.md

Required to appear. The root README.md is the project overview at /projects/<slug>. Relative links and images are rewritten to absolute GitHub URLs so they resolve off-site.

The README may carry optional YAML frontmatter to set metadata. The body below the frontmatter is the overview.

---
title: "zkRP: Encrypted P2P with Zero-Knowledge Range Proofs"
summary: "An encrypted peer-to-peer network built on zero-knowledge range proofs."
tags: [cryptography, p2p, reliability]
featured: true
slo: "99.9%"
sla: "99%"
---

# zkRP

The overview body starts here.

Frontmatter field	Effect	Falls back to
title	Project title	repo name
summary	One-line summary	repo description
tags	Tag pills	repo topics
featured	Featured on the home page	false
slo	Shows an SLO badge on the project page	none
sla	Shows an SLA badge on the project page	none

.nosite

Hides the repo. An empty file named .nosite at the root removes it from the site.

docs/

A documentation site. If the repo has docs/README.md (or docs/index.md), the whole docs/ tree is published at /projects/<slug>/docs with a searchable, nested sidebar. A folder becomes a nav section; its README.md/index.md is that section's page; each page's title is its first # H1.

docs/
  README.md
  guides/
    README.md
    setup.md

blog/

Articles that belong to the project. Drop Markdown files in blog/ and they are autoscanned, newest first. No blog/index.txt is needed. Each post uses frontmatter and appears in the site /blog feed and the home "In the pipeline" carousel.

---
title: "Building reproducible golden images"
date: 2026-07-18
summary: "One Packer pipeline, identical images across five clouds."
tags: [cloud]
---

Body.

Top-level blog/*.md are posts; blog/README.md and blog/index.md are ignored.

.resume.yml

A resume entry for the project. Indexed as the ultimate truth for that project and competes in the resume builder by score (give it a weight).

# .resume.yml  - omit or set `resume: false` to opt out
name: "Erid: Cloud Images"
period: "2026"
weight: 7            # base weight (1-10); competes by weight x recency
start: "2026-01"     # optional; drives the recency multiplier
tags: [cloud, sre]   # which resume presets this project appears on
strict: true         # optional; pin to ONLY the tagged presets (no borrowing)
tech: [Packer, Terraform]
bullets:
  - text: "Reproducible golden images across five clouds."
    tags: [cloud, sre]
    priority: 1

By default a project can be borrowed onto a roomier resume it is not tagged for, to help fill the page. Set strict: true to pin it to only its tagged presets, for example a project you want on the sales resume and nowhere else.

.related.txt

Links to related projects. One repo slug per line.

zkrp
erid

The one control in the site repo

.include_repo.txt (in the bangasser.dev repo) adds repos you do not own: put owner/name on its own line and that repo is fetched and shown like your own. This is the only project control that does not live in the project repo.