{"passport":{"unfragile":{"@version":"1.0","version":"2026-05","artifact":{"id":"github-hkuds--vibe-trading","slug":"hkuds--vibe-trading","name":"Vibe-Trading","type":"agent","url":"https://pypi.org/project/vibe-trading-ai/","page_url":"https://unfragile.ai/hkuds--vibe-trading","categories":["ai-agents"],"tags":["ai-agent","algorithmic-trading","backtesting","fintech","llm","mcp","multi-agent","python","quantitative-finance","trading"],"pricing":{"model":"open_source","free":true,"starting_price":null},"status":"active","verified":false},"capabilities":[{"id":"github-hkuds--vibe-trading__cap_0","uri":"capability://planning.reasoning.multi.agent.orchestration.for.trading.decisions","name":"multi-agent orchestration for trading decisions","description":"Coordinates multiple specialized AI agents (market analysis, risk management, execution) that communicate via MCP (Model Context Protocol) to collaboratively generate trading signals and validate decisions before execution. Each agent operates as an independent reasoning unit with access to shared market data and portfolio state, enabling distributed decision-making with built-in consensus mechanisms for trade approval.","intents":["I want multiple AI agents to analyze different aspects of a trade (technical, fundamental, risk) and reach consensus before executing","I need to decompose complex trading logic across specialized agents that can reason independently about market conditions","I want to implement a checks-and-balances system where one agent proposes trades and another validates risk parameters"],"best_for":["quantitative traders building multi-perspective trading systems","teams implementing AI-driven trading with explainable decision chains","developers prototyping collaborative agent architectures for financial decision-making"],"limitations":["agent communication overhead adds latency to trade execution — unsuitable for sub-millisecond HFT strategies","consensus mechanisms may reject valid trades if agents disagree, reducing trade frequency","no built-in agent failure recovery — if one agent crashes, the entire orchestration halts unless explicitly handled"],"requires":["Python 3.9+","MCP-compatible LLM provider (OpenAI, Anthropic, or local Ollama instance)","real-time market data feed (API key for broker or data provider)"],"input_types":["market data (OHLCV candles, order book snapshots)","portfolio state (positions, cash, P&L)","natural language trading rules or strategies"],"output_types":["structured trade signals (symbol, side, quantity, stop-loss, take-profit)","agent reasoning traces (decision logs from each agent)","execution status (filled, rejected, pending)"],"categories":["planning-reasoning","multi-agent-systems"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"github-hkuds--vibe-trading__cap_1","uri":"capability://data.processing.analysis.backtesting.engine.with.agent.replay","name":"backtesting engine with agent replay","description":"Simulates historical market conditions and replays trading agent decisions against past price data to evaluate strategy performance without risking capital. The engine reconstructs historical market state (OHLCV, order book), feeds it to agents in chronological order, and measures outcomes (Sharpe ratio, max drawdown, win rate) while preserving agent reasoning logs for post-hoc analysis and debugging.","intents":["I want to test my trading agent strategy against 5 years of historical data before deploying to live markets","I need to compare performance of different agent configurations (different LLM models, prompt variations) on the same historical data","I want to identify which market conditions cause my agent to make poor decisions by analyzing backtest logs"],"best_for":["quantitative traders validating strategies before live deployment","researchers benchmarking agent-based trading approaches","developers iterating on agent prompts and decision logic"],"limitations":["backtesting assumes perfect execution at historical prices — does not account for slippage, partial fills, or market impact from large orders","agent reasoning may differ in live markets due to latency, data freshness, or LLM non-determinism even with fixed seeds","memory-intensive for multi-year backtests with high-frequency data — requires significant disk/RAM for storing agent decision traces"],"requires":["Python 3.9+","historical OHLCV data (CSV, Parquet, or broker API with historical endpoints)","LLM API credentials (for replaying agent decisions during backtest)","minimum 2GB RAM for 1-year backtest with hourly data"],"input_types":["historical OHLCV candles (timestamp, open, high, low, close, volume)","agent strategy definition (prompts, decision rules, risk parameters)","portfolio initialization (starting capital, position limits)"],"output_types":["performance metrics (total return, Sharpe ratio, max drawdown, win rate, trade count)","trade-by-trade log (entry price, exit price, P&L, agent reasoning)","equity curve (portfolio value over time)","agent decision traces (reasoning from each agent for each trade)"],"categories":["data-processing-analysis","automation-workflow"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"github-hkuds--vibe-trading__cap_10","uri":"capability://memory.knowledge.agent.decision.logging.and.explainability","name":"agent decision logging and explainability","description":"Captures detailed logs of agent reasoning for every trade decision, including the market data considered, decision rules applied, and confidence scores. Enables post-trade analysis and debugging by providing full visibility into why agents made specific decisions, supporting both automated analysis and manual review.","intents":["I want to understand why my agent made a losing trade by reviewing its reasoning logs","I need to audit my agent's decisions for compliance purposes, showing the data and logic behind each trade","I want to identify patterns in agent decision-making that correlate with poor performance"],"best_for":["traders debugging agent behavior and improving strategies","teams implementing compliance-grade audit trails for regulated trading","researchers analyzing agent decision-making patterns"],"limitations":["detailed logging adds overhead to agent execution — can increase decision latency by 10-50% depending on logging verbosity","storing full decision logs for millions of trades requires significant disk space — requires log rotation or archival strategies","LLM reasoning logs may be difficult to parse and analyze programmatically — requires custom parsing logic"],"requires":["Python 3.9+","disk storage for decision logs (minimum 1GB per 10,000 trades with detailed logging)","optional: log analysis tools or dashboards for visualizing decision patterns"],"input_types":["agent decision (trade proposal with reasoning)","market data context (prices, indicators, news)"],"output_types":["decision log (timestamp, market data, decision rules applied, confidence score, final decision)","explainability report (human-readable explanation of decision rationale)","decision pattern analysis (clustering of similar decisions, correlation with outcomes)"],"categories":["memory-knowledge","safety-moderation"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"github-hkuds--vibe-trading__cap_11","uri":"capability://data.processing.analysis.live.market.sentiment.and.news.integration","name":"live market sentiment and news integration","description":"Integrates real-time market sentiment data (social media, news feeds, sentiment scores) and feeds it to agents as additional context for trading decisions. Agents can incorporate sentiment signals alongside technical and fundamental analysis to identify trades with higher conviction or avoid trades during negative sentiment spikes.","intents":["I want my agent to consider market sentiment from news and social media when making trading decisions","I need to avoid trading during periods of extreme negative sentiment that often precede market reversals","I want to identify trades with higher conviction by combining technical signals with positive sentiment indicators"],"best_for":["traders incorporating alternative data sources into algorithmic strategies","teams building sentiment-aware trading systems","researchers studying the impact of sentiment on agent trading behavior"],"limitations":["sentiment data quality varies significantly across sources — some sentiment scores are unreliable or manipulated","sentiment signals are often lagging indicators — by the time sentiment appears in data feeds, market may have already moved","integrating sentiment adds complexity to agent decision-making — may increase false signals if sentiment is overweighted"],"requires":["Python 3.9+","sentiment data provider subscription (e.g., Refinitiv, Bloomberg, alternative data vendors)","NLP capabilities for processing raw news/social media text (built-in or via external API)"],"input_types":["sentiment scores (numerical sentiment from -1 to +1 or categorical positive/negative/neutral)","news headlines and summaries","social media sentiment (Twitter, Reddit, etc.)"],"output_types":["sentiment-adjusted trade signals (trades with sentiment confirmation)","sentiment context (current sentiment score, recent sentiment trends)","sentiment-based risk adjustments (position sizing adjusted based on sentiment extremes)"],"categories":["data-processing-analysis","memory-knowledge"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"github-hkuds--vibe-trading__cap_2","uri":"capability://data.processing.analysis.real.time.market.data.ingestion.and.state.management","name":"real-time market data ingestion and state management","description":"Continuously fetches live market data (price ticks, order book updates, news) from broker APIs or data providers and maintains a synchronized portfolio state (positions, cash, P&L, margin) that agents can query. Implements connection pooling, automatic reconnection, and data validation to ensure agents always operate on fresh, consistent market state without stale data causing incorrect decisions.","intents":["I want my trading agents to always have access to the latest market prices and order book depth without manually polling APIs","I need to ensure my portfolio state (positions, cash) stays synchronized with the broker even if the connection drops and reconnects","I want to validate incoming market data for anomalies (price gaps, volume spikes) before feeding it to agents"],"best_for":["traders running live trading agents that require sub-second data freshness","teams building multi-asset trading systems with complex portfolio state","developers implementing resilient data pipelines for financial applications"],"limitations":["network latency and broker API rate limits introduce delays — real-time data may lag actual market by 100-500ms depending on broker","no built-in data persistence — if process crashes, historical tick data is lost unless explicitly logged to disk","broker API outages cause data gaps — agents may make decisions on stale data if fallback mechanisms are not implemented"],"requires":["Python 3.9+","broker API credentials (Interactive Brokers, Alpaca, Binance, etc.) or data provider subscription","network connectivity with <500ms latency to broker servers","minimum 512MB RAM for maintaining order book state for 10+ symbols"],"input_types":["broker API endpoints (WebSocket or REST for price/order book data)","portfolio state queries (positions, cash balance, margin requirements)"],"output_types":["normalized market data (symbol, timestamp, bid/ask, last trade price, volume)","portfolio state snapshots (positions, cash, unrealized P&L, margin utilization)","data quality metrics (freshness, gaps, anomalies)"],"categories":["data-processing-analysis","tool-use-integration"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"github-hkuds--vibe-trading__cap_3","uri":"capability://text.generation.language.natural.language.strategy.definition.and.interpretation","name":"natural language strategy definition and interpretation","description":"Accepts trading strategies written in natural language (e.g., 'buy when RSI drops below 30 and price is above 200-day MA') and converts them into executable agent prompts and decision rules via LLM interpretation. The framework parses strategy descriptions, extracts technical indicators and conditions, and generates agent instructions that can be executed against live or historical market data.","intents":["I want to describe my trading strategy in plain English without writing code and have it automatically converted to executable agent logic","I need to quickly test multiple strategy variations by tweaking natural language descriptions rather than refactoring code","I want to generate agent prompts from strategy descriptions that capture my trading intuition and risk preferences"],"best_for":["non-technical traders who want to implement AI-driven strategies without coding","quantitative researchers prototyping strategy ideas rapidly","teams collaborating on strategy design where non-engineers contribute strategy descriptions"],"limitations":["LLM interpretation of natural language is non-deterministic — same strategy description may generate slightly different agent prompts across runs","ambiguous or vague strategy descriptions may result in agent behavior that doesn't match trader intent","no validation that generated prompts are actually executable or profitable — requires manual backtesting to verify"],"requires":["Python 3.9+","LLM API credentials (OpenAI, Anthropic, or local Ollama)","basic understanding of trading concepts (indicators, risk management)"],"input_types":["natural language strategy description (plain text or structured template)","optional: reference to technical indicators (RSI, MACD, Bollinger Bands, etc.)"],"output_types":["agent prompts (system and user messages for LLM)","structured decision rules (entry conditions, exit conditions, position sizing)","risk parameters (stop-loss, take-profit, max position size)"],"categories":["text-generation-language","planning-reasoning"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"github-hkuds--vibe-trading__cap_4","uri":"capability://planning.reasoning.risk.management.and.position.sizing.with.agent.validation","name":"risk management and position sizing with agent validation","description":"Enforces portfolio-level risk constraints (max drawdown, max position size, max leverage) and validates proposed trades against these constraints before execution. Agents can propose trades, but a dedicated risk management agent evaluates each trade's impact on portfolio risk metrics and either approves, rejects, or modifies the trade to comply with risk parameters.","intents":["I want to ensure no single trade can cause my portfolio to exceed a 10% drawdown threshold","I need to automatically reject trades that would push my leverage above 2x, even if the trading agent proposes them","I want to dynamically adjust position size based on current portfolio volatility and margin availability"],"best_for":["traders managing significant capital who need strict risk controls","teams implementing compliance-grade risk management for regulated trading","developers building risk-aware multi-agent systems"],"limitations":["risk calculations assume normal market conditions — extreme volatility or market gaps can cause actual losses to exceed calculated VaR","position sizing based on historical volatility may be too conservative in trending markets or too aggressive in calm markets","no built-in integration with broker margin requirements — requires manual synchronization of risk parameters with broker limits"],"requires":["Python 3.9+","portfolio state with historical returns (for volatility calculation)","broker API for querying margin requirements and available buying power","risk parameter configuration (max drawdown %, max leverage, max position size %)"],"input_types":["proposed trade (symbol, side, quantity, entry price)","current portfolio state (positions, cash, margin utilization)","risk parameters (max drawdown, max leverage, max position size)"],"output_types":["trade approval/rejection decision","modified trade parameters (adjusted quantity to comply with constraints)","risk impact analysis (projected portfolio drawdown, leverage, margin utilization after trade)"],"categories":["planning-reasoning","safety-moderation"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"github-hkuds--vibe-trading__cap_5","uri":"capability://tool.use.integration.trade.execution.with.broker.integration.and.order.management","name":"trade execution with broker integration and order management","description":"Submits approved trades to brokers via their APIs, manages order lifecycle (pending, filled, partially filled, cancelled), and handles edge cases like order rejections, partial fills, and slippage. Maintains a mapping between agent-proposed trades and actual broker orders, enabling reconciliation and logging of execution outcomes for performance analysis.","intents":["I want to submit trades approved by my agents directly to my broker without manual intervention","I need to handle partial fills and adjust my portfolio state accordingly when orders don't fill completely","I want to track the difference between agent-proposed prices and actual execution prices to measure slippage"],"best_for":["traders automating trade execution across multiple brokers","teams implementing order management systems for algorithmic trading","developers building resilient execution layers that handle broker API failures"],"limitations":["broker API latency introduces execution delays — orders may fill at worse prices than proposed if market moves quickly","no built-in order cancellation logic — if an order partially fills and market conditions change, manual intervention may be needed","broker API rate limits may queue orders, causing delays in execution during high-volume trading periods"],"requires":["Python 3.9+","broker API credentials with trading permissions (not just read-only)","broker support for order types used by agents (market, limit, stop-loss, etc.)","network connectivity with <100ms latency to broker servers for time-sensitive orders"],"input_types":["approved trade from risk management agent (symbol, side, quantity, order type, price limits)","broker API credentials and account ID"],"output_types":["order confirmation (order ID, status, filled quantity, average fill price)","execution report (entry price, exit price, slippage, commission, net P&L)","order status updates (pending, filled, cancelled, rejected)"],"categories":["tool-use-integration","automation-workflow"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"github-hkuds--vibe-trading__cap_6","uri":"capability://data.processing.analysis.performance.analytics.and.strategy.evaluation","name":"performance analytics and strategy evaluation","description":"Calculates trading performance metrics (Sharpe ratio, Sortino ratio, max drawdown, win rate, profit factor) from trade logs and equity curves, enabling comparison of different agent configurations and strategies. Generates visualizations (equity curves, drawdown charts, trade distribution) and statistical summaries to help traders understand strategy behavior and identify improvement opportunities.","intents":["I want to compare the Sharpe ratio and max drawdown of two different agent configurations to decide which to deploy","I need to understand which market conditions my strategy performs well in and which cause losses","I want to visualize my equity curve and drawdown over time to identify periods of poor performance"],"best_for":["quantitative traders evaluating strategy performance","researchers benchmarking agent-based trading approaches","teams conducting post-trade analysis and strategy optimization"],"limitations":["metrics assume consistent market conditions — strategies optimized for one market regime may underperform in different regimes","Sharpe ratio and other backward-looking metrics don't predict future performance","visualization libraries may struggle with very large datasets (millions of trades) — requires data aggregation or sampling"],"requires":["Python 3.9+","trade log with timestamps, entry/exit prices, and P&L","optional: matplotlib or plotly for visualization"],"input_types":["trade log (timestamp, symbol, side, entry price, exit price, quantity, P&L)","equity curve (timestamp, portfolio value)"],"output_types":["performance metrics (Sharpe ratio, Sortino ratio, max drawdown, win rate, profit factor, average trade duration)","visualizations (equity curve, drawdown chart, monthly returns heatmap, trade distribution histogram)","statistical summaries (mean return, volatility, skewness, kurtosis)"],"categories":["data-processing-analysis"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"github-hkuds--vibe-trading__cap_7","uri":"capability://text.generation.language.agent.prompt.engineering.and.optimization","name":"agent prompt engineering and optimization","description":"Provides tools for iteratively refining agent prompts to improve trading performance, including prompt templates for common trading strategies, A/B testing framework for comparing prompt variations, and feedback mechanisms to help agents learn from past trades. Enables traders to experiment with different prompt phrasings, risk preferences, and decision-making styles without changing core agent logic.","intents":["I want to test how different prompt phrasings affect my agent's trading decisions and performance","I need to find the optimal balance between aggressive and conservative trading by tweaking agent prompts","I want to provide feedback to my agent about past trades so it can improve future decisions"],"best_for":["traders optimizing agent behavior through prompt engineering","researchers studying how LLM prompts affect trading decisions","teams iterating on agent strategies with limited code changes"],"limitations":["prompt optimization is empirical and non-deterministic — optimal prompts may vary across market conditions and LLM models","A/B testing requires running multiple agent instances simultaneously, increasing API costs and computational overhead","agent learning from feedback is limited by LLM context window — cannot retain all historical trades for long-term learning"],"requires":["Python 3.9+","LLM API credentials with sufficient quota for multiple concurrent agent instances","backtesting or paper trading environment for safe prompt experimentation"],"input_types":["base agent prompt (system message, decision rules, risk preferences)","prompt variations (alternative phrasings, different risk parameters)","trade feedback (which trades were profitable, which were losses, market conditions)"],"output_types":["optimized prompts (refined versions with better performance)","A/B test results (performance comparison across prompt variations)","prompt effectiveness metrics (correlation between prompt features and trading performance)"],"categories":["text-generation-language","planning-reasoning"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"github-hkuds--vibe-trading__cap_8","uri":"capability://planning.reasoning.multi.asset.and.multi.timeframe.strategy.support","name":"multi-asset and multi-timeframe strategy support","description":"Enables agents to simultaneously analyze and trade multiple assets (stocks, crypto, forex) across different timeframes (1-minute, hourly, daily) with coordinated decision-making. Agents can correlate signals across assets and timeframes to identify higher-conviction trades and avoid conflicting positions in correlated instruments.","intents":["I want my agent to trade both stocks and crypto simultaneously while avoiding overlapping positions in correlated assets","I need to combine signals from multiple timeframes (daily trend + hourly entry) to improve trade quality","I want to diversify across asset classes while maintaining coordinated risk management across the portfolio"],"best_for":["traders managing multi-asset portfolios with coordinated strategies","teams building diversified algorithmic trading systems","researchers studying cross-asset correlations in agent-based trading"],"limitations":["coordinating decisions across multiple assets and timeframes increases agent reasoning complexity and latency","correlation assumptions may break down during market stress — assets assumed uncorrelated may move together","data synchronization across multiple timeframes and assets adds complexity — requires careful handling of asynchronous data updates"],"requires":["Python 3.9+","market data for multiple assets and timeframes (requires multiple data subscriptions or unified data provider)","broker support for trading multiple asset classes simultaneously"],"input_types":["multi-asset market data (OHLCV for each asset and timeframe)","correlation matrix (historical correlations between assets)","portfolio constraints (max exposure per asset class, max correlation)"],"output_types":["coordinated trade signals (trades across multiple assets with correlation checks)","portfolio composition (allocation across assets and timeframes)","correlation analysis (actual vs expected correlations, correlation breakdowns)"],"categories":["planning-reasoning","data-processing-analysis"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"github-hkuds--vibe-trading__cap_9","uri":"capability://automation.workflow.paper.trading.and.live.trading.mode.switching","name":"paper trading and live trading mode switching","description":"Provides a unified interface for running agents in paper trading mode (simulated execution without real capital) and live trading mode (real execution with actual capital), with seamless switching between modes. Enables safe testing of agents on live market data before deploying real capital, with identical execution logic and risk management in both modes.","intents":["I want to test my agent on live market data without risking real capital before switching to live trading","I need to run my agent in paper trading mode for a week to validate performance before going live","I want to switch my agent from paper to live trading with a single configuration change, ensuring identical logic in both modes"],"best_for":["traders validating agents on live market data before deploying real capital","teams implementing gradual rollout strategies (paper → small live → full live)","developers testing agent behavior under realistic market conditions"],"limitations":["paper trading doesn't account for market impact or slippage from large orders — live performance may differ significantly","paper trading assumes perfect execution at market prices — doesn't simulate order rejections or partial fills","switching between paper and live modes requires careful state management — bugs in mode switching can cause unintended live trades"],"requires":["Python 3.9+","broker API credentials for both paper and live trading accounts","identical market data feed for paper and live modes (to ensure consistent agent inputs)"],"input_types":["mode configuration (paper vs live)","broker credentials for selected mode","agent strategy and risk parameters"],"output_types":["trade execution (simulated in paper mode, real in live mode)","portfolio state (simulated cash/positions in paper, real in live)","execution reports (identical format in both modes)"],"categories":["automation-workflow","tool-use-integration"],"confidence":0.5,"matches":0,"success_rate":0}],"trust":{"score":46,"verified":false,"data_access_risk":"low","permissions":["Python 3.9+","MCP-compatible LLM provider (OpenAI, Anthropic, or local Ollama instance)","real-time market data feed (API key for broker or data provider)","historical OHLCV data (CSV, Parquet, or broker API with historical endpoints)","LLM API credentials (for replaying agent decisions during backtest)","minimum 2GB RAM for 1-year backtest with hourly data","disk storage for decision logs (minimum 1GB per 10,000 trades with detailed logging)","optional: log analysis tools or dashboards for visualizing decision patterns","sentiment data provider subscription (e.g., Refinitiv, Bloomberg, alternative data vendors)","NLP capabilities for processing raw news/social media text (built-in or via external API)"],"failure_modes":["agent communication overhead adds latency to trade execution — unsuitable for sub-millisecond HFT strategies","consensus mechanisms may reject valid trades if agents disagree, reducing trade frequency","no built-in agent failure recovery — if one agent crashes, the entire orchestration halts unless explicitly handled","backtesting assumes perfect execution at historical prices — does not account for slippage, partial fills, or market impact from large orders","agent reasoning may differ in live markets due to latency, data freshness, or LLM non-determinism even with fixed seeds","memory-intensive for multi-year backtests with high-frequency data — requires significant disk/RAM for storing agent decision traces","detailed logging adds overhead to agent execution — can increase decision latency by 10-50% depending on logging verbosity","storing full decision logs for millions of trades requires significant disk space — requires log rotation or archival strategies","LLM reasoning logs may be difficult to parse and analyze programmatically — requires custom parsing logic","sentiment data quality varies significantly across sources — some sentiment scores are unreliable or manipulated","builder identity is not verified yet","no observed match outcomes yet"],"rank_breakdown":{"adoption":0.6057851163970357,"quality":0.34,"ecosystem":0.6000000000000001,"match_graph":0.25,"freshness":0.75,"weights":{"adoption":0.25,"quality":0.25,"ecosystem":0.1,"match_graph":0.28,"freshness":0.12}},"observed_outcomes":{"matches":0,"success_rate":0,"avg_confidence":0,"top_intents":[],"last_matched_at":null},"maintenance":{"status":"active","updated_at":"2026-05-24T12:16:21.550Z","last_scraped_at":"2026-05-03T13:57:04.027Z","last_commit":"2026-05-02T10:01:01Z"},"community":{"stars":4670,"forks":963,"weekly_downloads":null,"model_downloads":null,"model_likes":null}},"distribution":{"claim_url":"https://unfragile.ai/submit?claim=hkuds--vibe-trading","compare_url":"https://unfragile.ai/compare?artifact=hkuds--vibe-trading"}},"signature":"Lf8aQnaiW38NsYoc/xYjJW3CqQ4Yp52HFrwBsJfvqXxDUZLmYYurLNTmYnlaRSAYL/VLOwotMoRmzaL8bAt6Ag==","signedAt":"2026-06-21T17:10:36.894Z","signedBy":"unfragile.ai","version":1},"_links":{"self":"https://unfragile.ai/api/v1/passport/hkuds--vibe-trading","artifact":"https://unfragile.ai/hkuds--vibe-trading","verify":"https://unfragile.ai/api/v1/verify?slug=hkuds--vibe-trading","publicKey":"https://unfragile.ai/api/v1/trust-passport-public-key","spec":"https://unfragile.ai/trust","schema":"https://unfragile.ai/schema.json","docs":"https://unfragile.ai/docs"}}