Documents

EODHD CLI

Try it

Plan, build, review, or extend a stateless Bash CLI wrapper for the publicly documented EODHD API, intended for OpenClaw agents and Clawhub.ai distribution....

What it does

Build the wrapper as a small, stateless Bash CLI that translates stable subcommands into EODHD HTTPS requests and emits machine-friendly output. Keep secrets outside the repo and outside the CLI's own persistence model by aligning auth to OpenClaw-managed environment injection.

The skill document

EODHD

Overview

Build the wrapper as a small, stateless Bash CLI that translates stable subcommands into EODHD HTTPS requests and emits machine-friendly output. Keep secrets outside the repo and outside the CLI's own persistence model by aligning auth to OpenClaw-managed environment injection.

Workflow

  1. Confirm the task shape.

    • If the user is asking for a plan, load references/implementation-plan.md.
    • If the user is asking how to wire credentials safely, load references/openclaw-secrets.md.
    • If the user is asking for actual code changes, implement the narrowest useful slice first and preserve stateless behavior.
  2. Hold the contract steady.

    • Prefer eodhd as the binary name.
    • Default to JSON on stdout and diagnostics on stderr.
    • Exit non-zero on HTTP, auth, or argument errors.
    • Do not create config files, caches, token stores, or hidden state.
  3. Keep auth OpenClaw-native.

    • Expect EODHD_API_KEY to be injected by OpenClaw.
    • Allow a one-shot explicit override like --api-key only if the user asks for it or the implementation already exposes it.
    • Do not read plaintext secrets from repo files, sample configs, shell history helpers, or generated artifacts.
    • Keep the publish metadata aligned with that contract by declaring EODHD_API_KEY as the primary required credential.
  4. Enforce redaction and logging hygiene.

    • Never print the raw API key to stdout or stderr.
    • Redact secrets in debug output, dry-run output, request traces, and test snapshots.
    • Avoid set -x unless xtrace output is redirected through a sanitizer.
    • Treat query strings, headers, and copied curl commands as sensitive if they can contain api_token.
  5. Prefer a narrow, agent-friendly surface.

    • Start with stable commands such as eod, fundamentals, real-time, search, exchanges, and raw.
    • Add generic --query key=value support before adding many bespoke flags.
    • Use a raw escape hatch for unsupported endpoints instead of blocking new EODHD API coverage.
  6. Preserve public-distribution constraints.

    • Assume the repo is public.
    • Use only public EODHD docs and public sample symbols in examples.
    • Avoid local-machine assumptions, personal data, or environment-specific artifacts.
    • Verify the release bundle contains scripts/eodhd and the metadata that declares it.

Implementation Rules

  • Keep the CLI dependency-light: Bash plus curl as the baseline.
  • Use deterministic argument parsing and stable exit codes.
  • Separate URL construction, request execution, error translation, and output handling into small shell functions.
  • Make debug behavior opt-in with --verbose or --dry-run.
  • If tests are added, prefer smoke tests around URL generation, argument validation, and redaction.

Deliverables

  • scripts/eodhd for the CLI entrypoint when implementation starts
  • scripts/check-package.sh for package and manifest consistency checks
  • references/implementation-plan.md for milestones and acceptance criteria
  • references/openclaw-secrets.md for secret and logging constraints
  • agents/openai.yaml for Clawhub/OpenAI-facing metadata

Validation

  • Check that examples never expose a real token.
  • Check that debug or dry-run output replaces token values with a fixed mask such as ***REDACTED***.
  • Check that auth failure, HTTP failure, and usage failure produce distinct exit codes or clearly distinct stderr messages.
  • Check that the manifest declares EODHD_API_KEY and includes scripts/eodhd in the packaged files.
  • Check that the repo contains no .env, token snapshots, or generated local artifacts.

Related skills

Generate and edit Draw.io, Mermaid, and Excalidraw diagrams from natural language using a structured JSON spec.

by nssa.io1.0k installs47 stars

Stores durable facts in a categorized, plain-markdown vault on disk, alongside your agent's built-in memory.

by Iván559 installs18 stars

Find why your productivity system keeps failing, then apply the smallest fix — capacity math, bottleneck routing, durable local notes.

by Iván855 installs69 stars

Fetch raw ad creative, app, ranking, and revenue data from AdMapix as structured JSON.

by fly0pants4.3k installs296 stars

Join a video meeting as an AI bot with voice, avatar, and screenshare across four operating modes.

by johnpatternai22 installs8 stars

More from oscraters

Browse all skills

Use this skill to fetch TAAPI.IO indicator data for crypto or stocks, including fast single-indicator requests and bulk/multi-construct queries for agentic t...

by oscraters32 installs

Comprehensive Interactive Brokers (IBKR) TWS/Gateway skill using ib_insync. Includes Python and bash CLIs for account, market data, historical data, contract...

by oscraters44 installs1 stars

Use Weibo Open Platform for OAuth2 authentication, timeline retrieval, topic search, and structured social sentiment collection. Trigger this skill when task...

by oscraters48 installs1 stars

Agentic trading and market-data workflows for Futu OpenD (MooMoo/Futu OpenAPI), including OpenClaw-compatible secret-ref credential loading, account discover...

by oscraters38 installs

Intent-alignment orchestration for OpenClaw agent teams across diverse host environments. Use when work must stay anchored to user goals while allowing flexi...

by oscraters34 installs

Use this skill when users need Alpha Vantage market data or indicators (stocks, ETFs, forex, crypto, commodities, macro, company fundamentals) via the offici...

by oscraters53 installs