| name | tradecat-public |
| description | TradeCat public Hermes skill: Agent/Hermes paper-trading runtime using public online sheet signals, bundled Binance public-readonly resources, schema-audited context, and local paper/watch execution. |
tradecat-public Skill
Use this skill when Hermes/Agent needs to operate TradeCat from the embedded Skill package at skills/tradecat-public/. The Python project root is the repository root two levels above this Skill directory; the Skill package provides activation rules, Agent role profiles, references, and the canonical manifest.
When to Use This Skill
- The task mentions TradeCat public sheet signals,
signal_flow, anomaly_panel, Agent trading loops, paper/watch, paper ledger, replay/backtest, health/daily/alert reports, or autonomous paper ops.
- The task mentions Binance public/read-only Agent market context, K lines, order book, funding, open interest, long/short ratio, context audit, or trade thesis.
- The task needs the Agent machine contract, role profile, safety boundary, Hermes/OpenAI adapter config, or Skill/package governance.
Not For / Boundaries
- Do not read Binance keys, secrets, listen keys, account state, balances, positions, or orders.
- Do not sign requests, call private account/order/leverage/margin endpoints, or place real orders.
- Do not invent sizing, leverage, stop loss, take profit, or time-stop defaults; Agent thesis must explicitly provide them.
- Do not restore the retired local TUI, installer, watchdog, or cache-browser product line.
- Do not write runtime, cache, logs, ledgers, audit journals,
.env, .venv, .tradecat, .runtime, .hermes, or .tools into git.
- TradeCat may write local paper/watch state only under ignored
.runtime/.
