HK IPO Review Optimizer#

Use this skill for the post-decision feedback loop.

Runtime#

This publish bundle includes the required CLI runtime under runtime/hkipo-next.

From the skill folder:

bash

cd <skill_dir>
uv run --directory runtime/hkipo-next hkipo-next ...

By default review history is stored in ~/.hkipo-next/data/hkipo.db.

Workflow#

1. Use review list to find target records.
2. Use review show before updating a record.
3. Use review update to add actual results and variance notes.
4. Use review export when another tool needs a JSON dataset.
5. Use apply-suggestions show before accepting or rejecting imported suggestions.

Commands#

List recent review records:

bash

cd <skill_dir>
uv run --directory runtime/hkipo-next hkipo-next review list --limit 20 --format json

Show a record:

bash

cd <skill_dir>
uv run --directory runtime/hkipo-next hkipo-next review show rvw_123 --format json

Update actual results:

bash

cd <skill_dir>
uv run --directory runtime/hkipo-next hkipo-next review update rvw_123 \
  --allocated-lots 2 \
  --listing-return-pct 14.2 \
  --exit-return-pct 9.8 \
  --realized-pnl-hkd 1860 \
  --notes "Sold into first-day strength" \
  --variance-note "Grey market was weaker than expected but sponsor demand held" \
  --format json

Export a review dataset:

bash

cd <skill_dir>
uv run --directory runtime/hkipo-next hkipo-next review export --from 2026-04-01 --to 2026-04-16 --output /tmp/hkipo-review.json --format text

Preview and accept a suggestion:

bash

cd <skill_dir>
uv run --directory runtime/hkipo-next hkipo-next apply-suggestions show sgg_123 --format json
uv run --directory runtime/hkipo-next hkipo-next apply-suggestions accept sgg_123 --reason "Matches observed listing-day slippage" --format json

Output Cues#

• Review records preserve the original prediction payload, data status, and source issue count.
• Suggestion detail output shows whether a proposed change would create a new parameter version.

Companion Skills#

• Use $hkipo-parameter-manager when a review conclusion needs manual tuning work.
• Use $hkipo-decision-engine to rerun fresh decisions after a rule change.