Skip to content
sdlc

Goal-directed readiness and blocker introspection

Status: open/proposed

Statement

Section titled “Statement”

Beyond “what is runnable anywhere,” the orchestrator and its users want goal-directed readiness: the next N tasks that actually advance a chosen milestone, and — when nothing advances it — which unfinished prerequisites block it. The global frontier scatters effort; a goal-directed frontier focuses it and turns “why is this milestone stuck?” into a query.

Who/what it affects

Section titled “Who/what it affects”
  • Milestone owners tracking a specific deliverable to done.
  • /sdlc:orchestrate when it should concentrate dispatch on one milestone.
  • Anyone asking “what is blocking M-XXXX?” — a question with no home today.

Evidence

Section titled “Evidence”
  • Milestones aggregate an ordered set of member tasks (intersect’s M-0010 is described as “an ordered six-task set”); finishing a milestone means finishing its transitive task dependencies.
  • The orchestrator currently dispatches across all ready tasks with no milestone focus (solutions/ontological/skills/orchestrate/SKILL.md).
  • Blocker introspection is done ad hoc by reading depends_on chains by hand; nothing computes the blocking frontier.

Toward resolution

Section titled “Toward resolution”

PR-DZTZ-graph-scheduler’s goal-directed wrapper: the frontier intersected with ancestors(goal), plus a whatBlocks(goal) query returning the unfinished prerequisites not yet in the frontier. Same push engine, restricted to the goal’s subgraph — see D-BPD8-graph-scheduler-api.

← Back to Drivers