About this skill

Problem

When stock-scoring analysis is only a local script, it usually lacks a stable external execution path: calls need credentials, multi-turn sessions need context, long-running tasks need polling, and payment flows must return to the user. stock-scoring-analysis-v2 closes those gaps by wrapping interface setup, client execution, result display, and error checking into one run flow.

How It Works

Credentials first: check PRANA_SKILL_API_FLAG; if missing, fetch pk_...:sk_... from GET /api/v2/api-keys and write it to the environment.
Runtime selection: prefer Node.js 18+, with Node 20.10+ for ESM execution; fall back to Python 3 when Node is unavailable.
Session continuity: unless it is the first turn or -n starts a new session, include the previous data.thread_id; state files are preferred across independent OpenClaw processes.
Run the task: call POST /api/claw/agent-run with an x-api-key header, then deliver data.content to the user as-is, whether it is a URL or plain text.
Long tasks: if the response times out, returns 5xx, stays running, or is not valid JSON, query POST /api/claw/agent-result; the default waits 120 seconds before the first check and then polls every 120 seconds for up to 20 attempts, adjustable via PRANA_AGENT_RESULT_POLL_INTERVAL_SEC and PRANA_AGENT_RESULT_POLL_MAX_ATTEMPTS.

Boundaries

This skill fits teams that already have a Prana or OpenClaw environment and need to deliver stock-scoring output directly to engineers or business users. It does not define the scoring model, nor does it replace review, storage, or trading decisions. When payment is required, keep the returned payment link; for purchase history, send the full data.url to the user but avoid logging it.

Screenshots

Use Cases

Set up pk/sk auth in OpenClaw and run the Node client for stock scoring requests.
Pass the previous thread_id across multi-turn stock scoring calls to preserve context.
Poll agent-result every 120 seconds when agent-run stays running or times out.
Return the data.url from purchase-history-url directly to the user on request.

Best For

Platform engineers integrating Prana or OpenClaw: need to wire auth, Node execution, and polling into a system.
Product engineers wrapping stock scoring services: need to preserve thread_id turns and return data.content unchanged.
Full-stack engineers handling payments and order queries: need to retrieve purchase-history links without logging URLs.
Script maintainers using the Python fallback: need to run the Python client and inspect responses when Node is unavailable.

Related Skills

Cold Shower Gate

@user_2da54b45/cold-shower

Switch AI from a polite executor into a skeptical thinking partner, using assumption checks, pre-mortems, and evidence gates to expose weak ideas early.

Professional 2026.08.29 ↓ 280