agentsclimarketplace
Sign in

Relay claude

Skill yeaight7/agent-powerups/skills/relay-claude

Use when another local agent needs a one-shot Claude CLI answer without cross-turn relay memory.From its SKILL.md

Install
npx -y skills add yeaight7/agent-powerups --skill relay-claude

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 6 stars6 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

2.0 KB, 424 tokens by cl100k_base, as published. Nobody here has run it

Purpose

Send a bounded prompt to the local Claude CLI through Agent Powerups relay when Claude is useful as an external reviewer or advisor.

Compatible with: codex, gemini-cli, generic

Requires:

  • Local Claude Code CLI (claude)

When to Use

  • You need a second opinion from Claude on a focused question.
  • The task benefits from Claude's reasoning, but not from a persistent session.
  • You want an artifacted relay turn instead of pasting terminal output into chat.

Do not use for multi-turn advisory context. Use relay-gemini or a dedicated session workflow when context must persist.

Inputs

  • Relay session name
  • Focused prompt with repository context, constraints, and expected output
  • Optional files or snippets summarized into the prompt

Workflow

  1. Verify the CLI is available:
claude --version
  1. Start a relay session for Claude:
apx relay start <name> --provider claude
  1. Ask one focused question:
apx relay ask <name> "<prompt>"
  1. Read the relay artifact before acting on the advice.
  2. Stop the relay session when done:
apx relay stop <name>

Output

  • A relay artifact containing Claude's answer
  • A short local summary of whether the answer is actionable
  • Any follow-up questions captured as new, explicit relay asks

Verification

  • claude --version worked or missing CLI was reported.
  • Prompt included enough context for a one-shot answer.
  • Relay artifact was read before using the result.
  • Claude's advice was verified against repository files before edits.
  • Session was stopped or intentionally left active.

Failure Modes

  • Treating a one-shot Claude answer as verified fact.
  • Sending vague prompts that lack file paths, constraints, or output format.
  • Expecting cross-turn memory from the Claude relay.
  • Leaving relay sessions running without a reason.

What ships with it

Read from the repository

Just SKILL.md. No reference files, no scripts.

Keep looking

Skills are one crate of 326,401. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.