Run a broad project discovery pass because the planfile queue is idle.
Goal: move Koru from local implementation mode back to whole-project strategy,
then create concrete tickets and let the normal queue work through them.
-
Generate fresh whole-project context:
code2llm ./ -f all -o ./project --no-chunk --exclude '*.md'
-
Ask IDE LLM after the code2llm refresh:
"Co jeszcze zostalo do wykonania? zrob z tego nastepne tickety do planfile."
Keep the answer ticket-oriented (no broad multi-file edits in this step).
-
Convert findings into tickets:
koru scan --apply --semcod-artifacts --source koru-scan
-
Prefer high-signal work: failing gates, god modules, duplicated code, high
cyclomatic complexity, missing tests, and architecture seams that block future tickets.
-
When new tickets exist, stop broad discovery and work the tickets one by one.
After those tickets are done and the queue is empty again, another discovery pass
is allowed.
Strategy Metadata:
Run a broad project discovery pass because the planfile queue is idle.
Goal: move Koru from local implementation mode back to whole-project strategy,
then create concrete tickets and let the normal queue work through them.
Generate fresh whole-project context:
code2llm ./ -f all -o ./project --no-chunk --exclude '*.md'
Ask IDE LLM after the code2llm refresh:
"Co jeszcze zostalo do wykonania? zrob z tego nastepne tickety do planfile."
Keep the answer ticket-oriented (no broad multi-file edits in this step).
Convert findings into tickets:
koru scan --apply --semcod-artifacts --source koru-scan
Prefer high-signal work: failing gates, god modules, duplicated code, high
cyclomatic complexity, missing tests, and architecture seams that block future tickets.
When new tickets exist, stop broad discovery and work the tickets one by one.
After those tickets are done and the queue is empty again, another discovery pass
is allowed.
Strategy Metadata: