agentmods.dev

headless-relay

headless-relay is a skill for Claude Code, Codex from dorukardahan/headless-relay. It costs 252 tokens per session (31,091 once invoked), scanned D, original, MIT.

A guide for handing work from one coding agent to another AI model through command-line tools.

In plain words
What is it for?
Running GPT, GLM, Grok, Gemini, or Claude from an agent session and bringing their responses back for review.
Why use it?
It lets an agent request another model's opinion or output without leaving the current session.

Skill for Claude CodeCodex

Which agent this was written for is unclear — built for openclaw. Also seen: reads .claude/ paths; mentions subagents; positional $N argument.

Needs its repository: it reads a path above its own folder, which exists only inside the repository. The line is apd=$(cd "$(dirname "$ap")" 2>/dev/null && pwd -P) || { echo "grok_relay: auth directory not readable: $(dirname "$ap")" >&2; exit 1; } # PHYSICAL path: a sym.

not rated 1repo changed 10d ago D scan 252 tokens original MIT

Good fit Running GPT, GLM, Grok, Gemini, or Claude from an agent session and bringing their responses back for review.

Compare 6 skills from other repositories ↓
View source ↗ dorukardahan/headless-relay
Install

Getting it into your agent

It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.

Clone the repo
git clone --depth 1 https://github.com/dorukardahan/headless-relay
agentmods
npx agentmods add skills/dorukardahan/headless-relay/headless-relay

Made for: Claude Code, Codex.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for headless-relay

README.md
[![agentmods](https://agentmods.dev/badge/skills/dorukardahan/headless-relay/headless-relay/github.svg)](https://agentmods.dev/skills/dorukardahan/headless-relay/headless-relay)
Your own site
<a href="https://agentmods.dev/skills/dorukardahan/headless-relay/headless-relay"><img src="https://agentmods.dev/badge/skills/dorukardahan/headless-relay/headless-relay/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for headless-relay

Your own site · 80×15
<a href="https://agentmods.dev/skills/dorukardahan/headless-relay/headless-relay"><img src="https://agentmods.dev/badge/skills/dorukardahan/headless-relay/headless-relay.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 252 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 31,091 The whole file, excluding the scripts and references it only reads on demand.
Security scan D 3 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00252 $0.31091
Opus 5.5 $0.00101 $0.12436
Sonnet 5.5 $0.00050 $0.06218
Haiku 4.5 $0.00025 $0.03109

Measured 10d ago against content hash 7d836f508ec4, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade D, and why

headless-relay scanned grade D with 3 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 10d ago.

The scan reads SKILL.md. This mod also ships 12 executable files (scripts/catalog_watchdog.py, scripts/kimi_relay.py, scripts/print-model-catalog.sh, …), listed below but not scanned — reading those needs a real analyzer, not pattern matching.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Reads agent configuration directoriesmediumAgent snooping

.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.

`~/.codex/config.toml` (`sandbox_mode`, `approval_policy`, reviewer features), so headless

Recursive force deletehighDestructive command

rm -rf with a variable or a broad path is one typo away from removing the wrong tree.

trap 'rm -rf "$base" "${sbx:-}" 2>/dev/null' EXIT # armed BEFORE sandbox exists: nothing can leak

Runs shell commandslowCapability

Expected in a hook, worth knowing in a rule or an instructions file.

res = subprocess.run([PS, "-A", "-o", "pid=,pgid=,stat="], stdin=subprocess.DEVNULL,
SKILL.md · 1,386 lines

How it starts

The opening of the file, as written. The whole thing — 1,386 lines — stays where its author put it; the contents beside it link to each section on GitHub.

headless-relay

This skill provides instructions for delegating a task to another AI model without leaving the current agent session. The user says "ask Codex", "get GLM's opinion", "run this by Kimi"; the orchestrating agent (Claude Code, Codex CLI, or another harness) writes the prompt, runs the target model's headless CLI through its shell tool, reads the model's stdout, and summarizes the answer back. Follow these patterns exactly.

The six targets

Target CLI Headless entry point Auth / plan
GPT codex (OpenAI Codex CLI) codex exec ChatGPT plan or OpenAI API key (codex login)
GLM opencode, or zcode (ships inside the ZCode desktop app) opencode run / zcode --prompt Z.ai Coding Plan (API key or ZCode app login)
Kimi K3 kimi (Kimi Code CLI) kimi -p Kimi membership through native device-code OAuth (kimi login)
Grok grok (xAI Grok Build) grok -p — via the grok_relay helper (empty HOME + clean temp GROK_HOME so it can't read your other tools' config or grok's own global rules; see the next section) SuperGrok (grok login) or XAI_API_KEY
Gemini agy (Google Antigravity CLI; the standalone Gemini CLI is not a drop-in lane, see references/cli-reference.md) agy -p / --print Google account via the Antigravity app/CLI
Claude claude (Claude Code) claude -p, or the harness's native subagent Anthropic auth of the current session

Kimi text relay — isolate the working directory

Kimi print mode is agentic: it starts in the current directory, uses auto permissions, and cannot be combined with --plan. A text-only review therefore runs from a fresh empty non-git directory with an empty skills directory. This reduces accidental project access; it is neither full context isolation nor an OS sandbox. Opt-in required before the first call: tell the user that global instructions ($KIMI_CODE_HOME/AGENTS.md and ~/.agents/AGENTS.md), global MCP servers and hooks, inherited environment/endpoint overrides, and native session/log retention remain active. Hooks can execute commands; MCP can spawn processes or send context elsewhere. Real-home session history may retain prompts after the temporary cwd is deleted. Review those settings with the user and obtain explicit consent to this agentic exposure; if unsuitable, skip Kimi. Never read credential files as part of this review. Give native Kimi a real repository cwd only for an explicitly requested agentic task.

Read the full file on GitHub · 1,386 lines

Files

What ships with it

18 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 10d ago Changed · +78 lines 7d836f508ec4
  2. 12d ago Changed · +1 lines d8d77136a2af
  3. 28d ago Changed · +188 lines fd8c6e591d1f
  4. 29d ago Changed · +174 lines · +74 tokens per session 60dcc7bb4d62
  5. 1mo ago Changed · +39 lines · +13 tokens per session 7ecb93b5082e
  6. 1mo ago Changed · +10 lines 5217fec99064
  7. 1mo ago Changed · +4 lines 988dbad26bb8
  8. 1mo ago Changed · +2 lines 66739ffd27f7
  9. 2mo ago First seen · 890 lines · 165 tokens per session scan D 93a7044587dd

Subscribe to this mod's changes

headless-relay is a skill published in the GitHub repository dorukardahan/headless-relay (1 stars, last pushed 10d ago), licensed MIT. It adds 252 tokens to every session and 31,091 once invoked, about $0.0010 per session on Opus 5.5. A static security scan graded it D with 3 findings (reads agent configuration directories, recursive force delete, runs shell commands). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Building agentsClaude Codeagent-skillsai-agentscodexheadlessskills
Related

headless-relay compared with skills from other repositories

headless-relay next to 6 skills from other repositories
ModTypeStarsLast pushLicenceScanTokensWorks with
headless-relaydorukardahan/headless-relaySkill110d agoMITD252Claude Code, Codex
claude-md-improveranthropics/claude-plugins-officialSkill47ktodayApache-2.0A82Claude Code
agent-platform-model-registrygoogle/skillsSkill24kyesterdayApache-2.0A60Claude Code, Codex
gke-workload-securitygoogle/skillsSkill24kyesterdayApache-2.0A171Claude Code, Codex
gke-reliabilitygoogle/skillsSkill24kyesterdayApache-2.0A85Claude Code, Codex
google-cloud-solution-agentic-analytics-spark-knowledge-cataloggoogle/skillsSkill24kyesterdayApache-2.0A132Claude Code, Codex
glance-canonical-layoutsandroid/skillsSkill10kyesterdayApache-2.0A56Claude Code