gpd-compact-state

To archive STATE.md history, run the compact-state workflow to move older blocks, completed decisions, and resolved blockers to STATE-ARCHIVE.md. This preserves a complete historical record while keeping your active context lean for analysis.

You should compact your STATE.md file when it grows beyond your target line count. The compaction process triggers automatically once the active file exceeds the specified length, moving older entries to STATE-ARCHIVE.md to maintain a manageable context size.

When archiving STATE.md entries, cross-references are updated to ensure traceability with the archive. The workflow preserves recent session records in the active document while moving older historical blocks to STATE-ARCHIVE.md and maintaining links between the two files.

Yes, compacting STATE.md enforces state.json synchronization after the compaction process completes. This ensures that your structured state data remains consistent with the archived and active markdown entries following the workflow execution.

The best way to manage unbounded project memory in STATE.md is automated archiving. The compact-state workflow moves older decisions, blockers, and metrics to STATE-ARCHIVE.md, reducing active context size while retaining recent items for decision-making.

The compaction workflow archives older blocks, completed decisions, and resolved blockers, but retains recent session records and recent items in STATE.md. It targets entries eligible for archiving to reduce active context size without losing current project state information.