CH·02CLI reference

deeprelay skill

Install the deeprelay skill for coding agents (Claude Code, Codex, OpenCode)

Install the deeprelay skill for coding agents (Claude Code, Codex, OpenCode)

Synopsis

Teaches coding agents to use deeprelay.ai through this CLI. The skill is a set of operating rules plus reference files: what deeprelay is, how to read the account (plan quota including the rolling 7-day windows, credit balance, spend limits, usage, the request log, API keys), how to run chat, embeddings, image and video generation, how to point a coding agent at deeprelay with deeprelay setup, and how to read common API errors. install writes it where each agent discovers it: a ~/.claude/skills/deeprelay/ directory (SKILL.md plus references/) for Claude Code, the same directory layout under ~/.config/opencode/skills/deeprelay/ for OpenCode, and a managed block in ~/.codex/AGENTS.md for Codex. show prints every file install would write.

The skill carries guardrails the agent must follow: it confirms with you before any command that spends money (chat, embeddings, image, video) or changes your account or tool config (plans, payment methods, spend limits, API keys, webhooks, setup), it never treats a request made before you saw the price as approval, and it never prints your API key. The reference files carry the depth: account and billing detail, inference flags and the video lifecycle, and a symptom → cause → fix guide for quota (429), spend-limit (402) and upstream errors. The guardrails live only in the body, never in a reference file, so a reference file can be omitted without weakening them.

deeprelay skill commands are purely local (they write agent config files) and work before deeprelay login.

Examples

# Install for every coding agent found on this machine
deeprelay skill install

# Install for one agent explicitly
deeprelay skill install --agent claude
deeprelay skill install --agent codex
deeprelay skill install --agent opencode

# Review everything that would be installed
deeprelay skill show
deeprelay skill show --agent codex

For the full guide — what installs where, what the agent can do, and the built-in spending guardrails — see docs/agent-skill.md.

Options

  -h, --help   help for skill

Options inherited from parent commands

      --api-base string   API base URL (override with DEEPRELAY_API_BASE env) (default "https://api.deeprelay.ai/v1")
      --debug             Enable debug logging to stderr
      --no-preflight      Skip the plan/credit check before inference requests (override with DEEPRELAY_NO_PREFLIGHT env)
  -o, --output string     Output format: table|json (default table on TTY, json otherwise) (default "table")

SEE ALSO

← The deeprelay CLI