agentleFS
ray5273ray5273/kr-research-kit/skills/kr-stock-data-pack/SKILL.md

kr-stock-data-pack

Skill2 starsChanged 8 days agoMITView on GitHub

Gather structured fact packs for Korean stock research. Use when the user needs source-dated inputs such as price context, filings, results, guidance, revenue mix, balance-sheet facts, governance data, valuation inputs, chart inputs, outside-view inputs, or a reusable company fact pack before writing a Korean stock memo or event note. Do not use for the final narrative memo itself.

The file

Show its rules and text

We split this skill into the rules it gives an agent

22 rules. About 0 are common, 15 are distinctive, and 7 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈15

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 7

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use this skill to assemble the inputs that a downstream Korean stock memo needs.

    line 8
  2. Distinctive:

    Treat prices, disclosures, IR materials, guidance, and company news as time-sensitive.

    line 12
  3. Distinctive:

    If the downstream pack shape is not already defined, ask the user what they actually need from this pack: valuation, chart, peers, outside …

    line 13
  4. Distinctive:

    Prefer primary sources first: DART, KRX disclosures, company IR, official financial statements, and governance pages.

    line 14
  5. Very specific:

    If the user mainly needs exact filing extraction for quarterly, half-year, segment, customer-concentration, backlog, or contract-disclosure …

    line 15
  6. Very specific:

    When order-backlog-analysis.md and order-backlog-data.json exist, ingest their dated official/proxy basis and artifact paths into a …

    line 16
  7. Distinctive:

    When the downstream memo needs more than official company framing, gather dated outside-view inputs from sell-side, specialist media, or …

    line 17
  8. Very specific:

    If the company is export-driven and customer, geography, or product mix is not separately disclosed, ingest kr-trade-flow-analysis …

    line 18
  9. Distinctive:

    Keep the output structured.

    line 19
  10. Very specific:

    If the workspace is writable, write the pack to analysis-example/kr/<company>/data-pack.md when the user asked for a reusable artifact.

    line 20
  11. Distinctive:

    Start from the agreed security, mode, and time horizon.

    line 25
  12. Distinctive:

    Put the exact date next to each metric block when refresh cycles differ.

    line 31
  13. Distinctive:

    Use markdown tables or JSON-like blocks that the downstream writer can reuse without re-parsing.

    line 35
  14. Distinctive:

    Read references/workflow.md for the data-gathering checklist.

    line 37
  15. Distinctive:

    Read references/output-format.md for the default pack shape.

    line 38
  16. Distinctive:

    Read ../kr-stock-analysis/references/blended-source-notes.md when the downstream memo needs outside views alongside primary-source facts.

    line 39
  17. Distinctive:

    Keep conflicting dates visible instead of flattening them into one fake reference date.

    line 44
  18. Very specific:

    When naver-insights.md exists in the company output directory, ingest its per-post entries into ## External Views with Source role: …

    line 47
  19. Very specific:

    When trade-flow-analysis.md exists in the company output directory, ingest only the confidence summary and source-dated inference rows into …

    line 49
  20. Very specific:

    When order-backlog-analysis.md exists, ingest its Order-driven company classification, official total, year/undated split, final disclosed …

    line 50
  21. Distinctive:

    Do not convert a thin source set into a confident valuation call.

    line 51
  22. Distinctive:

    Ask a short user-need check before gathering broad data blocks when the must-have sections are still unclear.

    line 52

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Korean Stock Data Pack

Use this skill to assemble the inputs that a downstream Korean stock memo needs.

Quick Start

  • Treat prices, disclosures, IR materials, guidance, and company news as time-sensitive.
  • If the downstream pack shape is not already defined, ask the user what they actually need from this pack: valuation, chart, peers, outside views, governance, filing follow-up, or only a narrow subset.
  • Prefer primary sources first: DART, KRX disclosures, company IR, official financial statements, and governance pages.
  • If the user mainly needs exact filing extraction for quarterly, half-year, segment, customer-concentration, backlog, or contract-disclosure detail, use kr-stock-dart-analysis first and then pull the rest of the pack around it.
  • When order-backlog-analysis.md and order-backlog-data.json exist, ingest their dated official/proxy basis and artifact paths into a distinct backlog block. Do not flatten contract-maturity proxy amounts into official backlog.
  • When the downstream memo needs more than official company framing, gather dated outside-view inputs from sell-side, specialist media, or detailed independent analysis.
  • If the company is export-driven and customer, geography, or product mix is not separately disclosed, ingest kr-trade-flow-analysis artifacts as External Evidence or Revenue Mix Proxy; keep confirmed disclosures and trade-flow inference in separate rows.
  • Keep the output structured. The point is to gather clean inputs, not to write the final thesis.
  • If the workspace is writable, write the pack to analysis-example/kr/<company>/data-pack.md when the user asked for a reusable artifact.

Workflow

  1. Define the pack scope. Start from the agreed security, mode, and time horizon.
  2. Gather dated facts. Collect the latest price context, filings, earnings facts, governance facts, valuation inputs, and chart inputs. When filing precision is the bottleneck, anchor this step to a prior kr-stock-dart-analysis output.
  3. Gather outside views when useful. Capture the few outside takes that materially change the debate, and label whether each is sell-side, specialist media, or independent analysis.
  4. Keep dates visible. Put the exact date next to each metric block when refresh cycles differ.
  5. Separate facts from interpretation. Include only light context. Save the thesis and conclusion for kr-stock-analysis.
  6. Deliver a reusable pack. Use markdown tables or JSON-like blocks that the downstream writer can reuse without re-parsing.

Read references/workflow.md for the data-gathering checklist. Read references/output-format.md for the default pack shape. Read ../kr-stock-analysis/references/blended-source-notes.md when the downstream memo needs outside views alongside primary-source facts.

Operating Rules

  • Say directly when revenue mix, customer concentration, or a valuation metric is not separately disclosed.
  • Keep conflicting dates visible instead of flattening them into one fake reference date.
  • Distinguish verified facts from your inference.
  • Label outside inputs by source role and keep their verification status visible.
  • When naver-insights.md exists in the company output directory, ingest its per-post entries into ## External Views with Source role: independent analysis, Verification status: unverified (blogger), and the post URL in Source. Do not treat blogger claims as primary-source numbers.
  • When analyst-report-insight.md exists in the company output directory, ingest its broker-coverage rows into ## External Views with Source role: sell-side consensus, Verification status: sell-side (broker), the report's landing URL in Source, and the per-broker 1-liner as the takeaway. Also carry the consensus TP median and rating distribution into the valuation / view block as a labeled street-consensus reference — do not treat analyst TPs as primary filings.
  • When trade-flow-analysis.md exists in the company output directory, ingest only the confidence summary and source-dated inference rows into ## External Views or ## Revenue Mix Proxy with Source role: trade-flow proxy, Verification status: Trade Flow Inference, and the confidence grade. Do not promote trade-stat proxy claims to filing facts unless DART/IR explicitly confirms them.
  • When order-backlog-analysis.md exists, ingest its Order-driven company classification, official total, year/undated split, final disclosed year, coverage ratio, chart basis, and PNG path into ## Order Backlog Inputs. Preserve contract-disclosure-maturity-proxy and official-total-only labels exactly.
  • Do not convert a thin source set into a confident valuation call.
  • Ask a short user-need check before gathering broad data blocks when the must-have sections are still unclear.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.