Lifelines Survival Analysis

基于 lifelines 库提供生存分析与 Cox 比例风险建模能力，支持残差诊断、参数化回归模型自定义、时滞转化率分析及比例风险假设检验。

Security Scan

The skill claims to provide lifelines-based survival analysis but its instructions and artifacts are dominated by a finance/backtest/trading blueprint (ZVT) that asks the agent to run Python commands, install packages, and read/write local ZVT data — the declared purpose and runtime expectations do not align.

Purpose & Capability

Name/description promise: lifelines survival analysis and Cox modeling. Actual SKILL.md and supporting files are a Doramagic finance blueprint (finance-bp-126) that describes a full data->backtest->trading pipeline (ZVT), trading semantic locks, and use cases such as A-share backtests and order execution. This is a clear mismatch: survival-analysis functionality would not normally include trading execution, sell-before-buy semantic locks, or ZVT recorder preconditions.

Instruction Scope

Although instruction-only (no code), the SKILL.md tells the agent to run precondition Python commands (import zvt, run recorders), check/create ~/.zvt, and obey an execution protocol that requires reloading seed.yaml and enforcing many domain constraints/locks. Those instructions direct filesystem writes, package installs (pip install zvt), and could lead to network activity (recorders contacting data providers). They also include fatal trading semantics (e.g., execute sell before buy, next-bar execution) that go beyond pure statistical modeling and could produce trading actions if coupled to an execution adapter.

Install Mechanism

No declared install spec (instruction-only), which is lower risk in isolation. However SKILL.md requires 'Python 3.12+ with uv package manager' and preconditions instruct users/agents to run pip install zvt and run zvt.init_dirs; those are implicit install steps executed at runtime if the agent follows preconditions. Because installs are not declared formally, the skill relies on the agent to fetch/execute third-party packages (zvt and possibly provider recorders).

Credentials

The skill declares no required environment variables but references ZVT_HOME and provider integrations (eastmoney, joinquant, akshare, qmt) that commonly require credentials or config. Preconditions test/modify ~/.zvt and expect writable directories. The skill thus expects access to filesystem and possibly provider credentials while not declaring or requesting them explicitly — a mismatch that can lead to unexpected credential use or filesystem changes.

Persistence & Privilege

always:false and no explicit persistent installation of the skill — good. But the execution protocol and preconditions instruct package installation, directory initialization, and test file creation in ~/.zvt, meaning the agent may install packages and write persistent files during use. There's nothing forcing the skill to always be enabled, but it can perform system-level changes when invoked.

What to consider before installing

This skill's description says 'lifelines survival analysis' but the instructions and reference files are a finance/backtest/trading blueprint (ZVT) that will ask the agent to run Python commands, install packages (e.g., zvt), create and write to ~/.zvt, and interact with data provider recorders (which may need API keys). Before installing or running: 1) Confirm you actually want a ZVT-style backtest/trading assistant rather than a pure lifelines analysis tool. 2) Review references/LOCKS.md, seed.yaml, and SKILL.md to understand mandatory semantic locks (they can halt or change behavior) and preconditions. 3) Do not enable autonomous invocation for this skill if you intend to connect to real trading/execution adapters — run it in an isolated environment first. 4) If you proceed, run it in a sandbox or VM, and avoid supplying credentials until you verify which providers are used and why. 5) Because the skill does not declare required credentials but references services that often need keys, expect to be prompted for secrets during runtime — only provide those after careful review.

生存分析建模 (lifelines-survival-analysis)

基于 lifelines 库提供生存分析与 Cox 比例风险建模能力，支持残差诊断、参数化回归模型自定义、时滞转化率分析及比例风险假设检验。

Pipeline

data_collection -> data_storage -> factor_computation -> target_selection -> trading_execution -> visualization

Top Use Cases (19 total)

Cox Model Residual Analysis (UC-101)

Diagnosing Cox proportional hazards model fit by computing and visualizing martingale, deviance, and delta_beta residuals to identify outliers and inf Triggers: cox residuals, martingale residual, deviance residual

Time-Lagged Conversion Rate Analysis (UC-103)

Modeling marketing conversion rates where there is a time lag between initial contact and conversion event, requiring specialized survival analysis te Triggers: conversion rate, time-lagged, marketing

Piecewise Exponential Survival Models (UC-104)

Fitting piecewise exponential survival models that allow different hazard rates in different time intervals, useful when hazard is non-constant over t Triggers: piecewise exponential, varying hazard, breakpoints

For all 19 use cases, see references/USE_CASES.md.

Execute trigger: When user intent matches intent_router.uc_entries[].positive_terms AND user uses action verb (run/execute/跑/执行/backtest/fetch/collect)

What I'll Ask You

• Target market: A-share (default), HK, or crypto? (US stocks in ZVT are half-baked — stockus_nasdaq_AAPL exists but coverage is thin)
• Data source / provider: eastmoney (free, no account), joinquant (account+paid), baostock (free, good history), akshare, or qmt (broker)?
• Strategy type: MACD golden-cross, MA crossover, volume breakout, fundamental screen, or custom factor?
• Time range: start_timestamp and end_timestamp for backtest period
• Target entity IDs: specific stocks (stock_sh_600000) or index components (SZ1000)?

Semantic Locks (Fatal)

ID	Rule	On Violation
SL-01	Execute sell orders before buy orders in every trading cycle	halt
SL-02	Trading signals MUST use next-bar execution (no look-ahead)	halt
SL-03	Entity IDs MUST follow format entity_type_exchange_code	halt
SL-04	DataFrame index MUST be MultiIndex (entity_id, timestamp)	halt
SL-05	TradingSignal MUST have EXACTLY ONE of: position_pct, order_money, order_amount	halt
SL-06	filter_result column semantics: True=BUY, False=SELL, None/NaN=NO ACTION	halt
SL-07	Transformer MUST run BEFORE Accumulator in factor pipeline	halt
SL-08	MACD parameters locked: fast=12, slow=26, signal=9	halt

Full lock definitions: references/LOCKS.md

Top Anti-Patterns (15 total)

• AP-INSURANCE-001: Implicit numeric format assumptions without validation
• AP-INSURANCE-002: Triangle axis construction with invalid temporal ordering
• AP-INSURANCE-003: Cumulative/incremental triangle representation misuse

All 15 anti-patterns: references/ANTI_PATTERNS.md

Evidence Quality Notice

[QUALITY NOTICE] This crystal was compiled from blueprint finance-bp-126. Evidence verify ratio = 53.0% and audit fail total = 27. Generated results may have uncaptured requirement gaps. Verify critical decisions against source files (LATEST.yaml / LATEST.jsonl).

Reference Files

File	Contents	When to Load
references/seed.yaml	V6+ 全量权威 (source-of-truth)	有行为/决策争议时必读
references/ANTI_PATTERNS.md	15 条跨项目反模式	开始实现前
references/WISDOM.md	跨项目精华借鉴	架构决策时
references/CONSTRAINTS.md	domain + fatal 约束	规则冲突时
references/USE_CASES.md	全量 KUC-* 业务场景	需要完整示例时
references/LOCKS.md	SL-* + preconditions + hints	生成回测/交易代码前
references/COMPONENTS.md	AST 组件地图（按 module 拆分）	查 API 时

---

Compiled by Doramagic crystal-compilation-v6.1 from finance-bp-126 blueprint at 2026-04-22T13:01:02.914448+00:00. See human_summary.md for non-technical overview.