Compare selective and full plans over a recorded change corpus
Before making selection mandatory, the team needs a defensible way to find false negatives rather than celebrate shorter CI runs.
- Focused work estimate
- 6h + prerequisites
- Priority in the scenario
- High
- Engineering practice
- Experimental design · CI engineering
Estimated field mix
- Quality engineering50%
- Developer tooling50%
Field percentages are editorial estimates of the ticket's engineering focus. They total 100%; they are not measured time, proficiency scores, or ownership evidence.
Review it, then add it to your workspace.
The board opens an editable draft; nothing is saved until you confirm it. Sign-in and workspace permissions apply, and Demo boards remain ephemeral.
Project context
A fictional internal tools team maintains eighteen packages. Engineers currently run everything because the affected-package script occasionally misses downstream consumers.
Setup prerequisites
- Create a local six-package fixture with a dependency cycle, shared compiler configuration, and two example changes.
Preceding work
Complete these dependencies, or supply their agreed outputs before taking this ticket.
- BMONO-101 · Document task inputs beyond package dependencies
- BMONO-102 · Parse workspace manifests without executing package scripts
- BMONO-103 · Handle cycles in the downstream impact graph
- BMONO-104 · Resolve renamed and deleted files against both revisions
- BMONO-106 · Make cache keys include generator versions and relevant environment
- BMONO-105 · Explain why each selected command is necessary
- BMONO-107 · Detect stale plan execution after the worktree changes
- BMONO-108 · Bound parallel commands and stop cleanly on cancellation
Acceptance criteria
- Define a corpus covering root, package, rename, and deletion changes.
- Compare selected task outcomes with full-run outcomes under identical inputs.
- Block adoption on an unexplained excluded failure and record uncertainty.
Implementation constraints
- Report workload and machine limits; do not promise general speedups from local measurements.
Verification to include
- Detect an intentionally omitted shared input.
- Show a safe exclusion and publish reproducible comparison commands.
Deliverables
- Adoption assessment and mismatch report.
Rollout and recovery
Run in shadow mode for the corpus; retain a one-command full-run override.
Value of the work
For the engineer: Practice graph analysis, reproducible commands, and safe developer-tool failure modes.
For the team: Produce an inspectable CI selection plan that avoids unnecessary work without silently skipping consumers.
Evidence boundaries
Outcome Evidence: Tests, patches, and runbooks are requested deliverables. They become Outcome Evidence only through a qualified Mission and immutable Evidence IDs.
Ownership Evidence: Independent adaptation must be observed under a declared verification policy and cite immutable Evidence IDs. Completing a planning ticket establishes no Ownership Evidence.