Automations
Create and manage scheduled or price-triggered automations for financial research.
Skill requires user confirmation before creating automations and explicitly lists parameters to clarify (schedule, thread, instruction, delivery), demonstrating least privilege and user confirmation. However, data flow transparency is insufficient: no explicit scope of automation permissions, data storage, or sensitive data handling. No mention of dependency security or rollback. Source attribution unverified; publisher not verified, so trust deducted.
Instructions are clear and self-consistent, with detailed parameter tables, examples, and schedule format. However, no error handling, edge cases, or failure feedback. No tests or execution evidence, so reliability low.
Skill clearly describes applicable scenarios (scheduled tasks, price-triggered) and audience. Detailed examples and parameters make triggering conditions relatively clear. However, non-fit boundaries not stated, and price-triggered functionality heavily depends on overseas service (ginlix-data) raising reachability concerns for mainland users. Deduction applied.
Documentation is well-structured with clear tool descriptions, examples, and thread strategy. But lacking install/dependency notes, versioning, changelog, and known limitations. License Apache-2.0 present but no version or maintenance responsibility. Deduction applied.
Skill clearly describes creating and managing automations, use cases clear. However, actual execution results not verified, direct usability of output unknown. Price-triggered relies on specific backend, may not work standalone. Marginal value limited, so significant deduction.
Files contain CI workflows and test files, but tests are not for this skill path. No third-party execution evidence, key claims unverified. Low confidence, so low score.
- Price-triggered automations rely on real-time WebSocket feed from ginlix-data, which may be inaccessible from mainland China; verify network reachability.
- Skill lacks installation, dependency, and versioning info; user must verify compatibility.
- Documentation does not cover failure or error scenarios for automations; user should test.
What does this skill do, and when should you use it?
This skill provides three tools for creating and managing automations: list or inspect automations, create new scheduled or price-triggered automations, and update, pause, resume, trigger, or delete existing ones. It supports cron-based recurring schedules, one-time ISO datetime schedules, and trigger conditions based on real-time price movements (e.g., price above/below, percentage change). Automations can be configured to run in a fresh thread, a persistent thread, or the current conversation, and optionally deliver results via channels like Slack. Best practice is to confirm details with the user before creating an automation to ensure it runs as expected.
Provides three tools: check_automations lists or inspects a specific automation with execution history; create_automation creates a new automation with time-based (cron or ISO datetime) or price-triggered conditions (e.g., price above, percentage change), plus thread strategy and delivery options; manage_automation supports update, pause, resume, trigger, or delete operations. Price triggers can be one-shot or recurring with a cooldown (min 4 hours) or default once per trading day.
- An investor wants a daily market briefing on weekdays at 9 AM, automatically summarizing overnight moves and news for their watchlist.
- A trader wants an immediate alert when AAPL drops below $150, with automatic analysis of recent news and sentiment.
- A portfolio manager wants a trigger when SPX moves +2% from yesterday's close, running an analysis of volume and fundamentals.
- A research analyst wants a weekly portfolio review every Monday at 9 AM, using a persistent thread to maintain context across runs.
- A crypto investor wants a warning when BTC moves more than 3% from today's open, with a 4-hour cooldown to avoid spam.
What are this skill's strengths and limitations?
- Supports multiple schedule formats (cron, ISO datetime, price-triggered) for flexibility.
- Price triggers support multiple condition types (price above/below, percentage change) and reference baselines (previous close, day open).
- Thread strategy options (new, persistent, current) accommodate different workflows.
- Optional delivery methods (e.g., Slack) for sharing results.
- Comprehensive management tools: pause, resume, trigger, delete.
- Price-triggered automations rely on real-time data source (ginlix-data), which may be limited to the hosted platform during beta; self-hosting may require additional setup.
- Documentation does not specify resource consumption or concurrency limits for automations.
- Requires confirmation before creation, adding an extra step for users.
- Price trigger conditions are limited to a single symbol; no multi-symbol combinations.
How do you install this skill?
This skill is part of the LangAlpha repository (ginlix-ai/LangAlpha), which bundles 32 skills. Clone the repository and deploy LangAlpha; the skill lives at skills/automation/SKILL.md. Follow the repository README for installation: run make config to set up, then make up to start services.
How do you use this skill?
The skill loads automatically within LangAlpha conversations. Provide a clear creation request, e.g., "Create an automation named 'AAPL Price Alert' that triggers when AAPL drops below $150, one-shot." or "Set up a morning market brief automation that runs on weekdays at 9 AM." The system will guide you through confirming schedule, thread strategy, instruction, and delivery.