Skip to content

Team Adoption Guide

This guide walks your team through adopting the FCC Agent Team Framework over three weeks, from foundational concepts through hands-on simulation to full customization.

Define Team Roles

Before starting, assign these responsibilities within your team:

Role Responsibility Typical Owner
FCC Lead Owns persona configuration, workflow selection, and quality gate thresholds Tech lead or documentation lead
Simulation Operator Runs simulations (deterministic and AI-powered), reviews trace output Developer or QA engineer
Governance Reviewer Reviews quality gate results, manages tag assignments, approves publications Senior engineer or compliance officer
Content Owner Reviews and approves generated documentation artifacts Subject matter experts
Integration Engineer Manages CLI tools, template customization, CI/CD integration DevOps or platform engineer

Not every role needs a dedicated person. On smaller teams, the FCC Lead and Simulation Operator are often the same individual.

Three-Week Training Plan

Week 1: Concepts and Orientation

Goal: Everyone understands FCC fundamentals and can read persona specifications.

Day Activity Resources
Mon Introduction to FCC: problem, solution, three phases What is FCC?
Tue Persona deep-dive: R.I.S.C.E.A.R. specification walkthrough Understanding Personas
Wed FCC cycle walkthrough: Find, Create, Critique phases The FCC Cycle
Thu Read through core persona YAML files in data/personas/ Core personas: RC, BC, DE, RB, UG
Fri Team discussion: map current documentation pain points to FCC capabilities Capability Mapping

Week 2: Hands-On Simulation

Goal: The Simulation Operator can run simulations and the team can interpret results.

Day Activity Resources
Mon Install framework, set up development environment First Simulation
Tue Run GEN-001 (Full FCC Cycle) in deterministic mode Scenario data/scenarios/starter_scenarios.json
Wed Examine trace output, understand message flow and coverage Reading Results
Thu Run extended workflow (20-node), compare trace complexity Multi-Persona Pipelines
Fri Team review: discuss workflow selection for your first real project Workflow Selection

Week 3: Customization and Deployment

Goal: The team can create custom personas, modify templates, and configure governance.

Day Activity Resources
Mon Create a custom persona YAML for a team-specific role Custom Personas
Tue Modify Jinja2 templates to match team documentation standards Custom Templates
Wed Configure quality gates and severity thresholds Custom Gates
Thu Write a custom validation scenario for a real project Custom Scenarios
Fri First sprint planning: assign personas, select workflow, set quality targets This guide

First Sprint Checklist

Use this checklist when launching FCC on your first real project:

  • Workflow selected: Base (5-node), Extended (20-node), or Complete (24-node)
  • Personas assigned: Map team members to persona review responsibilities
  • Quality gates configured: Thresholds set for your project's quality requirements
  • Scenario defined: Initial input, start node, expected artifacts documented
  • Template review: Jinja2 templates checked against team style guide
  • Deterministic dry run: Simulation completed in mock mode before AI-powered execution
  • Trace review: Output traces validated by Governance Reviewer
  • Publication channels: SCP persona configured with access controls and channel formatting

Measuring Adoption Success

Track these metrics over your first month:

Metric Target How to Measure
Simulation completion rate 100% of scenarios pass SimulationResult.pass_rate
Quality gate pass rate > 90% first-pass Count gates passing on first run
Documentation cycle time 30% reduction Compare pre/post FCC production time
Governance compliance 100% artifacts tagged Audit tag coverage via tag registry
Team confidence All roles operational Weekly self-assessment

Common Adoption Pitfalls

  1. Starting with the Complete (24-node) workflow -- Begin with Base (5-node) to build familiarity before scaling up.
  2. Skipping deterministic simulation -- Always validate workflows in mock mode before running AI-powered simulations.
  3. Ignoring feedback edges -- The critique-to-find and critique-to-create feedback loops are where quality improvements happen. Do not disable them.
  4. Under-configuring quality gates -- Default thresholds are starting points. Tune them to your project's specific requirements.

Next Steps