Skip to content

FCC for Open-Source Maintainers

Welcome to the track for open-source maintainers. This is written for people who run their own OSS projects — large or small, solo or team — and want to adopt some of FCC's patterns for their own governance, docs, and persona-based contribution workflows.

FCC is itself an open-source project (MIT license, public on GitHub). It has a few patterns that translate well beyond AI workflows: the R.I.S.C.E.A.R. specification as a general contributor-role template; docs-as-code with auto-generated cross-references; release-cadence rhythm from the v1.0 → v1.7 arc; sister-project coordination without merging codebases.

What you will learn

  • How to adapt R.I.S.C.E.A.R. as a contributor-role specification for your project.
  • A docs-as-code setup adapted from FCC's Jinja2 + cross-reference pattern.
  • A contributor-ceremony pattern that scales from solo to multi-maintainer.
  • Release-cadence lessons learned across FCC's 20+ releases.

How this track is organized

Page Focus Effort
Getting Started Orientation; adapt nothing yet 20 min
Adopting R.I.S.C.E.A.R. Contributor-role spec for your project 1-2 hours to draft
Docs-as-Code Patterns Build a docs pipeline from FCC's 3-4 hours
Contributor Ceremony Onboarding, review, release rituals 1-2 hours
Release Cadence Lessons What FCC learned the hard way reading
Further Reading ADRs, references, tools reference

Prerequisites

  • You run or help run an open-source project. Doesn't matter the language.
  • You have some experience with Markdown, YAML, and at least one template engine (Jinja, Handlebars, Liquid, etc.).
  • You use Git and GitHub (or a similar forge).

You do not need to be using FCC itself. This track is about adopting patterns from FCC.

Why bother?

Three reasons open-source maintainers tend to benefit from these patterns:

  1. R.I.S.C.E.A.R. is a remarkable role-description format. It makes implicit expectations explicit. Adapting it to define contributor roles ("Bug-triager," "Docs reviewer," "Release manager") tends to reduce onboarding time measurably.
  2. Auto-generated cross-references. FCC generates 56 doc files per persona with automatic cross-reference. A similar pattern applied to modules, packages, or APIs reduces drift between the code and the docs.
  3. Release hygiene. The FCC 1.x series (roughly monthly releases for two years) has surfaced patterns worth stealing: how to version, when to branch, when to patch, and how to document.

Cost of adoption

Honest accounting:

  • A couple of hours to read this track.
  • A day to draft your own role specs (adapted R.I.S.C.E.A.R.).
  • A day or two to set up a docs-as-code pipeline if you don't have one already.
  • Ongoing discipline to actually use what you built. The patterns pay off only if you use them.

What FCC is, in one paragraph (so you can orient non-AI patterns)

FCC is a Python framework for structured multi-agent AI workflows. Each "agent role" is defined in a ten-field YAML spec called R.I.S.C.E.A.R. Runs produce machine-readable traces. The code is under src/fcc/, docs under docs/, data (YAML personas, JSON schemas, workflows) under src/fcc/data/. It has 20+ releases, a plugin system, multiple sister projects, and ten constellation-vertical repos maintained by the same group.

You do not need to understand the AI parts to benefit from the documentation and governance patterns.

Non-AI projects that have adopted similar patterns

Three anonymized sketches:

  • A Rust library maintainer adapted R.I.S.C.E.A.R. to describe module ownership. Each module has a YAML file with role, constraints, responsibilities, and role_collaborators (other modules). CI enforces that every constraint in each module's YAML has a matching test.
  • A docs-site maintainer built a 3-template docs-as-code pipeline modelled on FCC's 23-template system. Reduced time to add a new doc page from "45 minutes" to "5 minutes."
  • A small nonprofit OSS adopted FCC-style ADRs for their decisions. After six months, new contributors cited the ADRs as their single most useful onboarding material.

What not to adopt

Not every FCC pattern generalizes:

  • 164 personas is a domain-specific number. Do not port it literally.
  • The compliance framework is specific to AI regulation.
  • The workflow graphs are specific to the Find/Create/Critique cycle.
  • The plugin system is useful only if your project has a genuine extension surface.

Copy patterns; do not cargo-cult them.

Where next

If you are unsure whether this is worth your time, go to Getting Started for a shorter orientation. If you know you want to try, start with Adopting R.I.S.C.E.A.R. — it is the pattern with the highest return for most projects.

Related reading in this repo: ADR-0001 Plugin System, ADR-0002 Persona Schema.