gza-summary

Gza Summary

Produce an operator triage summary by combining dedicated gza surfaces, then recommend the next concrete actions.

Process

Step 1: Gather task data

Run these commands to collect the canonical source data:

Failed-task history:

uv run gza history --status failed

Unimplemented plan/explore work:

uv run gza advance --unimplemented

Until uv run gza unimplemented exists, uv run gza advance --unimplemented is the canonical surface.

Unmerged code work:

uv run gza unmerged

Queue state, including blocked rows:

uv run gza next --all

Optional failed-recovery decision surface:

uv run gza watch --restart-failed --dry-run

Step 2: Analyze the output

Treat each command as the authoritative surface for one domain:

From uv run gza history --status failed:

• Treat this as factual failed-task history, including attempts that were later retried or resumed successfully
• Suggest factual follow-up like uv run gza log <full prefixed task id> when operators need details
• Use uv run gza watch --restart-failed --dry-run when operators need the current failed-task recovery decision surface
• Do not treat this as a replacement for the old mixed-bucket gza incomplete output

From uv run gza unmerged:

• Each listed row is completed code work that still needs merge/review/sync attention
• Prefer gza-native guidance like uv run gza merge <full prefixed task id> or uv run gza sync <full prefixed task id>

From uv run gza advance --unimplemented:

• Identify completed plan or explore rows that do not yet have implementation work
• Suggest bare uv run gza implement <full prefixed task id> only for completed plan rows; note that it queues by default
• Suggest uv run gza advance --unimplemented --create when implement tasks should be queued from listed source rows

From uv run gza next --all:

• Identify the runnable next task(s) for uv run gza work
• Call out blocked rows, including dropped-dependency blockers, as queue state rather than lineage state