ForgeSDLC
Navigate
Home
Discover ForgeSDLC (101)
Practice (201)
Master (301)

Forge — connection to the SDLC foundation

This blueprint treats Forge as a delivery methodology that plugs into a shared foundation: tracking (spine + artifacts) and ceremony intents (C1–C6). Forge does not replace the foundation; it defines how the team refines, challenges, executes, and releases work.

1. SDLC phases A–F (how Forge maps)

Phase Forge expression Notes
A — Discover Ore intake; discover: Sparks New ideas, problems, and opportunities enter the backlog as Ore
B — Specify Ore → Ingot refinement; specify: Sparks Ingots answer: what, why, constraints, evidence of done
C — Design Ingot decomposition into Sparks; design: Sparks Architecture, ADRs, technical planning
D — Build Charge execution; build: Sparks Daily work toward the iteration increment
E — Verify Assay Gate; verify: Sparks Evidence gathering: tests, reviews, acceptance
F — Release Assay Gate pass → release; release: Sparks Ship when evidence is sufficient; learning feeds new Ore

Prescriptive rule: A Forge iteration should touch A through F but not necessarily in sequence. Sparks from different phases coexist in a single Charge. Phase A (Ore intake) and F (learning) are continuous.

Forge teams still maintain the blueprint tracking spine:

Artifact Forge mapping
Intent / request Ore — raw, unrefined input from any source
Spec Ingot — refined work with acceptance criteria; may link to ADR/spec docs
Plan Sparks decomposed from Ingots; Charge is the daily plan
Tasks Sparks (the finest executable grain — same hierarchy level, richer contract)
PRs Implementation slices for build: and verify: Sparks
Reviews Versona challenges + peer review + Assay Gate evidence
Release Assay Gate pass; ship decision based on evidence

Prescriptive rule: Every Spark that ships should be linkable to its parent Ingot and the original Ore. The Ember Log captures decision context that links and commits alone cannot convey.

3. Ceremony intents (C1–C6) ↔ Forge ceremonies

Intent Primary Forge ceremony Secondary
C1 — Align & decide Refinement (Ore → Ingot), Versona challenge Stakeholder sessions, hat-switching declarations
C2 — Plan the slice Planning (Ingot → Sparks), Charge selection Iteration scope negotiation
C3 — Execute & unblock Daily sync (Charge confirmation) Ad-hoc pairing, Versonas for blocked Sparks
C4 — Review & quality Review (evidence assessment) Versona challenges at decision points
C5 — Reflect & improve Retro (learning → new Ore) Ember Log review, Versona reliability assessment
C6 — Assure / release Assay Gate (evidence-based release decision) Continuous via quality gates in CI

See ceremony foundation and methodology bridge.

4. Role archetypes (blueprint hats on a Forge team)

Forge hat Typical archetype emphasis Notes
Product Sponsor proxy + Orchestrator (prioritization) Owns Ore intake and Ingot acceptance; decides value
Engineering Implementer (primary) + Quality advocate Builds Sparks; owns technical quality and DoD
Challenge Quality advocate (cross-cutting) Versona agents + human challenge; surfaces blind spots
Governance Orchestrator (process) + Quality advocate (compliance) Assay Gate steward; release evidence; Ember Log discipline

Detail: roles-archetypes.md, Forge roles chapter.

5. What Forge adds beyond the foundation

  • Work state precision: Ore → Ingot → Spark → Charge (not just "to do / doing / done").
  • Explicit challenge: Versona agents linked to discipline knowledge bases and bridge documents.
  • Decision memory: Ember Log captures why at decision points.
  • Strategic pause: Banked vs Blocked distinction prevents hidden prioritization.
  • Evidence-based release: Assay Gate with per-work-type evidence requirements.
  • Hat-switching protocol: Explicit role transitions for solo and small teams.
  • Phase-tagged Sparks: discover:, specify:, design:, build:, verify:, release: prefix convention.

6. Anti-patterns (prescriptive "don't")

Anti-pattern Fix
Committing Ore directly to execution Enforce refinement: Ore must become Ingot before planning
Sparks too large (multi-day) Split: a Spark should complete in one focused session (1–4 hours)
Versonas as bureaucracy (challenge everything) Time-box Versona; use at decision points, not on routine work
Ember Log as status report Keep it concise: decisions, trade-offs, assumptions — not narration
Skipping Assay Gate under schedule pressure Assay Gate is non-negotiable; adjust scope, not evidence standards
Hiding prioritization as "blocked" Use Banked for strategic pause; Blocked is for external impediments only
Duplicating the WBS with Forge IDs Spark IDs inherit from existing Task IDs — one hierarchy, one namespace
Adding ceremony for levels that already fit If Stories already have acceptance criteria, they are Ingots — skip the extra step
Creating a separate Charge board Charge is a view (label/filter) on the existing backlog, not a new surface
Manual journal entry for every field Automate from git + Cursor tracking; contributor confirms hat and DoD only

7. References in-repo