Market Morning Brief

Daily morning and evening intelligence digest for prediction market traders. Morning brief: Kalshi portfolio P&L, Polymarket trending markets, crypto prices...

What to check before installing: - Secrets & config: The skill expects Kalshi credentials (api_key_id + a private_key_file) and optionally Coinbase keys — these are provided via a config.yaml (e.g., ~/.openclaw/config.yaml), not declared env vars. Inspect your config file for other credentials before the skill reads it. If you don't want it to access account keys, do not populate the Kalshi/Coinbase sections. - Local files it will read/write: It reads cache files under ~/.openclaw/state (or configured cache_paths) and will write evening_news_history.json there. If you use the same state directory for other skills, this skill can read caches produced by them. - Dynamic import risk: The script tries to import code from a 'kalshalyst' scripts directory by manipulating sys.path (e.g., import trade_ledger). That means if other skill code is present on disk, this skill may execute it. If you are concerned, review the skill's import lines and any referenced third-party skill code before running. - LLM & local services: The evening news mode relies on a Qwen model via Ollama (local runtime). The troubleshooting steps assume you run Ollama locally. The skill also contains robust JSON-parsing utilities to handle LLM outputs. - Optional external endpoints: The skill can send a Slack notification via OPENCLAW_SLACK_WEBHOOK (it reads an env var or config key). If you don't want outbound notifications, leave that unset. - Practical steps: run the scripts in a sandbox or inspect the files first. Search for functions named load_config or the kalshi/coinbase client initialization in the code to see exactly what files are read; consider running with --dry-run/--debug first and with no credentials populated. If you share a system with other OpenClaw skills, consider isolating skill state directories to avoid unintended code execution/credential access. Confidence explanation: The functionality is coherent with the described purpose, but multiple small mismatches (undeclared credential access via config, an environment variable used but not declared, and dynamic imports that can execute other skills' code) raise moderate concern; additional info (confirm where load_config reads config from and confirm whether the skill automatically reads ~/.openclaw/config.yaml without explicit user opt-in) would increase confidence.

Type: OpenClaw Skill Name: market-morning-brief Version: 1.1.5 The Market Morning Brief skill is a legitimate intelligence tool for prediction market traders. It consolidates data from Kalshi and Coinbase APIs, Polymarket, and news sources using a two-stage AI filtering pipeline (via local Ollama/Qwen). While the skill handles sensitive financial data and includes a Slack notification feature (morning_brief.py), these behaviors are clearly documented, user-configured, and essential to the stated purpose. The code follows good practices such as graceful degradation, demo modes for first-run safety, and avoids shell execution when calling external processes.