Hk Stock To Gougoubi Prediction

Turn Hong Kong stock research into Gougoubi-ready prediction proposals with deterministic market wording, deadline selection, resolution rules, and tags. Use...

Security Scan

The skill is an instruction-only bridge that formats Hong Kong stock analysis into Gougoubi prediction payloads and its declared inputs/behavior are consistent with that purpose.

Purpose & Capability

The name/description match the runtime instructions: the skill selects prediction candidates from an hk-stock-predictor subskill and formats a deterministic Gougoubi payload handed to gougoubi-create-prediction. The required subskills are appropriate for the stated task.

Instruction Scope

SKILL.md limits actions to running the listed subskills, filtering/selecting prediction candidates, and formatting fields (marketName, deadlineIsoUtc, rules, tags). It explicitly forbids creating markets unless the user requests actual creation. It does not instruct the agent to read unrelated files, environment variables, or exfiltrate data.

Install Mechanism

This is instruction-only (no install spec) which minimizes risk. INSTALL.md mentions running a local install script and a GitHub installer command, but no install scripts are present in the provided file list and the registry entry has no automated install steps — this is an operational inconsistency (not a security signal) worth verifying before following INSTALL.md commands.

Credentials

The skill declares no required environment variables, credentials, or config paths. Its behavior (formatting market payloads and calling subskills) does not require secrets. Note: actual network access or credential needs would come from the referenced subskills (hk-stock-predictor and gougoubi-create-prediction), so those should be audited separately.

Persistence & Privilege

The skill does not request always:true and is user-invocable only. It does not claim to modify other skills or system-wide settings. No elevated persistence or privileges are requested.

Assessment

This skill is instruction-only and appears coherent for converting HK stock analysis into Gougoubi-ready proposals. Before installing or using it: 1) verify the two required subskills (hk-stock-predictor and gougoubi-create-prediction) because those will perform network calls or require credentials — their behavior determines real risk; 2) don’t blindly run the INSTALL.md shell commands unless you inspect the referenced scripts (they are not present in the package); 3) confirm you want the agent to submit markets before asking it to create them (the skill warns not to submit unless the user requests it). If you need higher assurance, request the implementations / provenance of the two subskills and any network endpoints they call.

Use hk-stock-predictor to generate prediction candidates, then convert the best candidate into a Gougoubi-ready proposal and pass it to gougoubi-create-prediction.

Minimal Input

{
  "symbol": "00700",
  "horizon": "30d"
}

Theme-first input is also valid:

{
  "theme": "南向资金持续流入的港股通互联网龙头",
  "horizon": "14d"
}

Required Subskills

Use these skills in order:

1. hk-stock-predictor
2. gougoubi-create-prediction

If hk-stock-predictor is missing, do not continue to market creation. Ask the agent to perform the HK stock analysis workflow first.

Deterministic Workflow

1. Analyze the HK stock or theme.
   • Run the hk-stock-predictor workflow.
   • Collect predictionCandidates and recommendedPrediction.
2. Pick one prediction candidate.
   • Prefer the recommended candidate unless it violates the filters below.
3. Validate prediction quality.
   • Must be binary, range, event, or relative-performance.
   • Must have a concrete deadline.
   • Must have a public resolution source.
   • Must not rely on subjective judgment.
4. Convert to Gougoubi proposal fields.
   • marketName
   • deadlineIsoUtc
   • rules
   • tags
5. Pass the resulting market to gougoubi-create-prediction.
   • Minimal required user-facing fields remain marketName and deadlineIsoUtc.
   • The rest of the content should be generated deterministically from the chosen prediction.
6. Return both layers of output.
   • analysis summary
   • chosen prediction
   • Gougoubi payload
   • create transaction result if submission is requested

Market Filters

Only create markets that satisfy all of these:

• time-bounded
• externally resolvable
• single thesis, not multiple stacked conditions
• sufficiently liquid underlying or high-quality event source
• wording is concise and non-ambiguous

Do not create markets for:

• micro-cap or illiquid names likely to be manipulated
• vague sentiment questions
• open-ended macro narratives without a measurable threshold
• markets that cannot be resolved from a public page or published report

Gougoubi Conversion Rules

1. marketName

Use a short title with one measurable claim.

Good:

• Will 00700 close above HK$520 on 2026-04-30?
• Will 09988 outperform HSTECH between now and 2026-05-15?
• Will 02318 report YoY net profit growth above 10% in its next earnings release?

2. deadlineIsoUtc

Choose the earliest deadline that still allows clean resolution:

• close-price market: use market close date plus a short safety buffer
• earnings market: use expected report date plus a safety buffer
• relative-performance market: end date at the comparison window close

Default timezone for rules: Asia/Hong_Kong, unless the data source clearly reports in another timezone.

3. rules

Rules must explicitly define:

• resolution source
• metric field
• comparison operator
• timezone
• tie handling
• fallback source if primary source is temporarily unavailable

Use this template:

Resolution source:
- Primary: [source]
- Fallback: [fallback source]

Resolution rule:
- This market resolves YES if [exact measurable condition].
- This market resolves NO otherwise.

Measurement details:
- Timezone: Asia/Hong_Kong
- Observation time: [exact time or reporting event]
- Metric: [close price / YoY growth / index-relative return / etc.]
- Tie handling: equality resolves [YES or NO], as stated in market title interpretation.

4. tags

Always include:

• hong-kong-stocks
• one sector tag
• one catalyst tag
• one horizon tag
• when possible, align with project-known category ids such as finance, earnings, and tech

Examples:

• hong-kong-stocks
• finance
• internet
• earnings
• 30d

Output Schema

{
  "ok": true,
  "analysis": {
    "symbol": "00700",
    "companyName": "Tencent Holdings",
    "recommendedPrediction": {
      "title": "",
      "deadlineIsoUtc": "",
      "confidence": 0
    }
  },
  "gougoubiPayload": {
    "marketName": "",
    "deadlineIsoUtc": "",
    "rules": "",
    "tags": []
  },
  "createInput": {
    "marketName": "",
    "deadlineIsoUtc": ""
  },
  "createResult": null,
  "warnings": []
}

Response Format

# 港股到 Gougoubi 预测题

## 标的结论
- 标的:
- 核心判断:
- 时间窗口:

## 选中的预测题
- 题目:
- 截止时间:
- 置信度:
- 公开裁决来源:

## Gougoubi 参数
- marketName:
- deadlineIsoUtc:
- tags:

## 规则草案
[rules text]

## 提交结果
- status:
- txHash:
- proposalAddress:

Boundaries

• Do not submit on-chain or through the UI unless the user clearly wants actual creation.
• If only asked for preparation, stop after generating the Gougoubi-ready payload.
• Surface uncertainty if earnings date or source availability is unclear.
• Prefer event-driven or close-price markets over loose valuation narratives.

Additional Resource

• For a concrete Tencent example, see examples.md.