Acknowledgments

Specflow builds on the work of others. This page credits the projects and people whose ideas were adapted, extended, or inspired features in Specflow.


Table of contents

  1. forge by Ikenna N. Okpala
    1. Specific Adaptations
  2. V3 QE Skill by Mondweep Chakravorty
    1. Adaptation in Specflow
  3. Timebreez Project
    1. What Timebreez Validated
  4. What Is Original to Specflow
  5. Academic Foundations

forge by Ikenna N. Okpala

Repository: github.com/ikennaokpala/forge

forge is an AI-assisted development framework that introduced several patterns Specflow has adapted for its contract-driven workflow.

Specific Adaptations

forge Concept Specflow Adaptation
Security quality gates SEC-001 through SEC-005 default contract rules. forge’s approach to embedding security checks into the development loop was adapted into Specflow’s YAML contract format with forbidden_patterns and required_patterns. See Security & Accessibility Gates.
Accessibility quality gates A11Y-001 through A11Y-004 default contract rules. forge’s accessibility checks were translated into Specflow contract patterns that run as part of the standard contract test suite.
Self-healing fix loops The heal-loop agent. forge’s concept of automated violation repair was adapted into Specflow’s confidence-tiered fix pattern system with score evolution and the Platinum/Gold/Silver/Bronze tier model. See Self-Healing Fix Loops.
Single-file SKILL.md packaging The concept of packaging an entire agent’s capabilities into a single markdown file informed Specflow’s agent prompt format (scripts/agents/*.md).
No-mock testing philosophy TEST-001 and TEST-002 contract rules. forge’s stance against mock-heavy tests influenced Specflow’s default test integrity contracts that flag excessive mocking.

V3 QE Skill by Mondweep Chakravorty

The TinyDancer model routing pattern from the V3 QE Skill introduced the idea of routing different agent tasks to different model tiers based on task complexity.

Adaptation in Specflow

Specflow’s Model Routing system assigns each of the 23+ agents to a recommended model tier (Haiku, Sonnet, or Opus), producing a 40-60% reduction in token costs. The routing table and tier definitions were adapted from the TinyDancer pattern.


Timebreez Project

The Timebreez project served as Specflow’s primary production validation environment. Over 280+ GitHub issues were delivered using Specflow’s wave execution system, providing the real-world feedback that shaped every feature.

What Timebreez Validated

  • Wave orchestration at scale (40+ issue backlogs)
  • Three-tier journey gates in practice
  • Agent Teams mode with persistent teammates
  • Contract completeness enforcement
  • Parallel execution savings (3-5x faster than sequential)

What Is Original to Specflow

While Specflow draws on the work above, the following features are original contributions:

Feature Description
Contract YAML schema The specific YAML format for feature and journey contracts, including forbidden_patterns, required_patterns, scope negation, severity levels, and compliance checklists
Journey verification hooks Git hooks that automatically extract issue numbers from commits, look up journey contracts, and run only the relevant E2E tests
Wave orchestration The 8-phase pipeline (analyze, plan, spawn, execute, test, validate, close, report) with dependency-based parallel execution
23+ agent library The full set of specialized agents, each with defined triggers, inputs, outputs, and quality gates
Agent Teams mode Persistent peer-to-peer teammate coordination via Claude Code TeammateTool API, with issue-lifecycle, db-coordinator, quality-gate, and journey-gate agents
Three-tier journey gates Tier 1 (issue), Tier 2 (wave), Tier 3 (regression) enforcement with baseline management and defer journal
Contract completeness enforcement CI gate that verifies CONTRACT_INDEX.yml stays in sync with contract files on disk
DPAO methodology Discovery, Parallel, Analysis, Orchestration – the methodology for parallel agent execution

Academic Foundations

Specflow also builds on established computer science research. See Background & Academic Foundation for the full intellectual lineage, including:

  • Design by Contract (Meyer, 1986)
  • Property-Based Testing (QuickCheck, Claessen & Hughes, 2000)
  • Behavior-Driven Development (North, 2003)
  • Continuous Delivery (Humble & Farley, 2010)

View on GitHub