Bug
The current Data Machine worktree fails Homeboy's sealed source-package check because the repository exceeds the configured 4 MiB source bound.
Reproduction
homeboy source package check --path /path/to/data-machine
Observed:
source package exceeds configured bounds: measured 382 entries / 4199432 bytes; limits 1024 entries / 4194304 bytes
The repository is 5,128 bytes over the byte limit. Largest measured entries are docs/CHANGELOG.md (272,432 bytes), composer.lock (132,288 bytes), and inc/Abilities/AgentAbilities.php (117,863 bytes).
Expected
A clean Data Machine checkout should satisfy the source-package policy used for bounded Homeboy/Lab handoff, either by reducing included source size or by explicitly correcting the package policy/exclusions at the owning layer.
Notes
This was discovered while validating a net-negative production adapter contraction for #3332. The contraction itself reduces the measured tree and all repository package contract smokes plus the production ZIP build pass; this issue tracks the independent package-bound baseline failure rather than working around it in that PR.
Bug
The current Data Machine worktree fails Homeboy's sealed source-package check because the repository exceeds the configured 4 MiB source bound.
Reproduction
homeboy source package check --path /path/to/data-machineObserved:
The repository is 5,128 bytes over the byte limit. Largest measured entries are
docs/CHANGELOG.md(272,432 bytes),composer.lock(132,288 bytes), andinc/Abilities/AgentAbilities.php(117,863 bytes).Expected
A clean Data Machine checkout should satisfy the source-package policy used for bounded Homeboy/Lab handoff, either by reducing included source size or by explicitly correcting the package policy/exclusions at the owning layer.
Notes
This was discovered while validating a net-negative production adapter contraction for #3332. The contraction itself reduces the measured tree and all repository package contract smokes plus the production ZIP build pass; this issue tracks the independent package-bound baseline failure rather than working around it in that PR.