Skip to content

B-ZEZR-clear-pipeline-pattern-for-readiness-and-verification-checks

Status: closed/abandoned

Several multi-step processes in SDLC each roll their own sequencing, gating, and reporting today — most visibly task readiness (task-ensure-ready) and work verification (the verify / review flows), but also orchestrate ticks, entity audits, and the (planned) index regeneration. Each is a bespoke chain of “do step, check result, branch on pass/fail, report,” with its own conventions for how a stage reports, how a failure short-circuits, and how partial results surface.

Establish one named pipeline pattern these processes adopt: an explicit, composable structure of stages, each with declared inputs/outputs, a clear pass/fail/skip gate, and uniform reporting — so that readiness checks, verification checks, and similar multi-step flows share a single abstraction instead of reinventing the control flow.

  • Consistency — a reviewer (human or LLM) reads any check the same way: same stage/gate/report shape, same failure semantics.
  • Composability — stages become reusable across processes (a “frontmatter validates” gate is the same whether it runs in readiness, audit, or CI).
  • Determinism — a defined pipeline pattern is the natural seam for the deterministic-tail / LLM-head split (S-0004-sdlc-cli-llm-head-deterministic-tail, P-0001-prefer-deterministic-over-llm): stages are deterministic; only the judgment stages call an LLM.
  • Observability — uniform per-stage reporting makes multi-step runs legible and operable (P-0002-operationalize-activity).
  • Is this a Decision (a pattern/standard) and/or a runner-registry concern (D-VSLI-distributed-work-runner-architecture)?
  • Scope of “pipeline”: just readiness + verification, or a general primitive for any multi-step process (orchestrate, audit, regen)?
  • Relationship to the existing skill-step conventions vs a new first-class pipeline abstraction.

Closed 2026-08-16: superseded by D-VSLI-distributed-work-runner-architecture. E4’s process-registration contract and E5’s DAG format are the reusable pipeline pattern this asked for.


← Back to Backlog