HK IPO Snapshot Extractor

Extract a structured Hong Kong IPO snapshot with company fields, offer mechanics, timing, source provenance, and field-level quality signals. Use when the user asks about one IPO symbol, wants normalized facts, or needs machine-readable IPO data before scoring.

Use this skill for read-only single-symbol analysis.

Runtime

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

From the skill folder:

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

Workflow

1. Run snapshot for the target symbol.
2. Use --format json when another skill will consume the result.
3. Check issues, quality.missing_fields, quality.conflicts, and quality.overall_confidence before making strong claims.
4. Build human summaries from the normalized fields instead of paraphrasing raw upstream pages.

Commands

Machine-readable snapshot:

cd <skill_dir>
uv run --directory runtime/hkipo-next hkipo-next snapshot 2476 --format json

Readable markdown snapshot:

cd <skill_dir>
uv run --directory runtime/hkipo-next hkipo-next snapshot 2476 --format markdown

Export a snapshot file:

cd <skill_dir>
uv run --directory runtime/hkipo-next hkipo-next snapshot 2476 --format markdown --output /tmp/2476-snapshot.md

Output Cues

• field_sources shows which upstream source won per field.
• quality.conflicts exposes alternate values when sources disagree.
• meta.degraded or meta.data_status=partial usually means the snapshot is usable with caveats, not unusable.

Companion Skills

• Use $hkipo-parameter-manager before scoring if the user wants to tune rules.
• Use $hkipo-decision-engine when the user wants a participation recommendation.