market-analyze

name	market-analyze
description	Assess trend, regime, and risk state for one A-share symbol. Use when user wants current market interpretation from the normalized snapshot without action sizing, execution planning, or thesis narration.
argument-hint	["symbol"]
allowed-tools	Bash(python3 *), Read, Glob

Market Analyze

Assess scored market state for: $ARGUMENTS

Overview

• Implementation status: code-backed
• Local entry script: <bundle-root>/market-analyze/run.py
• Primary purpose: turn normalized market data into a readable current-state interpretation
• Research layer: descriptive analysis only
• Workflow stage: stage 4 Signal and State Interpretation
• Local executor guarantee: compute and return the current market snapshot summary, including score, bias, trend, regime, key levels, notes, and risk flags

Use When

• The user wants current trend, regime, support, resistance, and risk-state interpretation.
• The user wants descriptive market-state context before moving into decision-support, strategy-design, or technical-scan.
• The user needs a bounded summary of the current snapshot rather than a thesis memo or full report.

Do Not Use When

• The user wants a full report with decision and strategy sections. Use market-brief.
• The user wants a direct trading action or quantity. Use decision-support.
• The user wants technical-only pattern language or broader chart commentary beyond the runtime snapshot summary. Use technical-scan.
• The user wants raw normalized data only. Use market-data.
• The user wants broader catalyst, thesis, or variant-view analysis. Use analysis.

Inputs

• Normal case: one stock symbol.
• Optional --csv PATH: use a local CSV instead of the default market source.
• Optional --start, --end, --source: constrain the data-loading path.
• If symbol is omitted, the skill may reuse last_symbol from the same execution context.
• Data note:
  • the local runtime derives its state from the normalized daily-bar frame
  • the user should not assume an intraday, benchmark-relative, or multi-factor attribution engine exists here

Execution

Step 1: Confirm the descriptive boundary

Use market-analyze for current-state interpretation only. It describes what the normalized snapshot currently looks like; it does not approve a position, design an execution plan, or validate a thesis.

Step 2: Run the local executor

python3 <bundle-root>/market-analyze/run.py <symbol> [--csv PATH]

Step 3: Deliver the snapshot honestly

The local output should be interpreted as a bounded state summary built from the enriched market frame. The agent should state:

• the data source and date window
• whether the symbol was explicit or reused from session context
• that score and bias summarize current state heuristically
• that the skill does not imply portfolio action, catalyst interpretation, or execution readiness

Output Contract

• Minimum local executor output: human-readable text beginning with 市场分析.
• Core fields: symbol, board, score, bias, trend, regime, support, resistance, notes, and risk flags.
• Side effects: updates session memory for the current execution context.
• Caller-facing delivery standard:
  • identify the data source and date window
  • treat score and bias as descriptive state summaries, not expected-return claims
  • keep the answer interpretive and current-state oriented
  • say explicitly when the result does not incorporate fundamentals, catalysts, portfolio constraints, or execution assumptions

Failure Handling

• Parse and argument errors: non-zero exit with a readable 命令错误 message.
• Data-loading or normalization errors: readable failure text beginning with 执行失败:.
• Missing symbol with no reusable session symbol: readable guidance instead of a traceback.
• If the available data window is too thin for robust interpretation, say so directly rather than pretending high confidence.

Key Rules

• Keep this skill interpretive rather than prescriptive.
• Use it as the clean descriptive anchor for technical-scan, decision-support, and strategy-design.
• Do not let a scored snapshot summary masquerade as a trading instruction.
• Do not imply indicators or controls that the current runtime does not actually expose in this skill’s local output.

Composition

• Usually follows market-data or shares the same upstream loaders.
• Often feeds technical-scan, decision-support, strategy-design, and market-brief.