multi-harness

A coordinator for sending limited coding tasks to other local coding-agent programs, called harnesses, such as Pi, OpenCode, Codex CLI, Claude Code CLI, and DHS.

In plain words
What is it for?
Delegating explicitly requested work to another coding-agent program, comparing models or harnesses, and handling requests that require an external runtime.
Why use it?
It helps when a requested model or runtime is available only through another command-line tool, while keeping the task boundaries and results under review.

Skill for Claude CodeCodex

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/lftpadilla/agent-dev-kit/multi-harness
Any agent
npx skills add LFTPadilla/agent-dev-kit --skill multi-harness
Clone the repo
git clone --depth 1 https://github.com/LFTPadilla/agent-dev-kit

Made for: Claude Code, Codex.

Per session 75 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,900 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
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 $0.00075 $0.01900
Opus 5 $0.00037 $0.00950
Sonnet 5 $0.00015 $0.00380
Haiku 4.5 $0.00007 $0.00190

Measured 2d ago against content hash 3e2666c8448e, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

multi-harness scanned grade A with 0 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 2d ago.

The scan reads SKILL.md. This mod also ships 1 executable file (scripts/delegate.py), 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.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

overnight-task-kit/skills/multi-harness/SKILL.md · 140 lines

How it starts

The opening of the file, as written. The whole thing — 140 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Multi Harness (Universal Harness Adapter)

Coordinate local agent harnesses without losing control of scope, safety, or verification.

Use this skill as the primary agent. You remain the orchestrator: decide what to delegate, run the external harness, inspect its output and any file changes, then synthesize the result for the user.

Routing Gate

Apply this gate before diagnostics, profile selection, or running delegate.py:

  1. Default to the current harness's native subagents when they expose the requested model. Requests for generic subagents, workers, parallelism, or orchestration are not external-harness requests.
  2. Use this skill only with an explicit external signal: Pi, OpenCode, Codex CLI, Claude Code CLI, DHS, another/external/cross harness, multiple distinct harnesses, harness comparison, or an external-only runtime the user has asked to use.
  3. When the user explicitly requests an external runtime or specific CLI model, use the Universal Harness Adapter (delegate.py).
  4. If this skill was auto-triggered without an explicit external signal, do not run diagnostics or delegation scripts. Continue using the current harness's native subagent tools.
  5. If the requested harness or model is not available locally, report the limitation and ask for confirmation before switching to an alternative model, harness, or native subagents. Never silently substitute.

Examples:

  • "Spawn three Codex subagents" -> use native Codex subagents when exposed; if unavailable, report the limitation and confirm before using the Codex CLI profile.
  • "Parallelize the review with subagents" -> native subagents; do not use multi-harness.
  • "Delegate this review to Pi with GLM or DeepSeek" -> use multi-harness.
  • "Run this implementation via DeepSeek Harness (DHS/DSH) or OpenCode" -> use multi-harness.

Universal CLI Harness Execution

The adapter supports running bounded tasks across:

  • Codex CLI: codex exec --ephemeral [-m <model>] [-C <dir>] [--yolo]
  • Claude Code CLI: claude -p "<prompt>" [--dangerously-skip-permissions]
  • DeepSeek Harness (DHS / DSH): dsh exec --dir <dir> [--model <model>] [--yolo] (Headless execution engine; no terminal TUI)
  • Pi: pi --print --mode text [--model <model>] --tools <tools>
  • OpenCode: opencode run --dir <dir> [--model <model>] [--auto]

Read the full file on GitHub · 140 lines

Files

What ships with it

5 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. 2d ago First seen · 140 lines · 75 tokens per session scan A 3e2666c8448e

Subscribe to this mod's changes

multi-harness is a skill published in the GitHub repository LFTPadilla/agent-dev-kit (2 stars, last pushed 4d ago), licensed MIT. It adds 75 tokens to every session and 1,900 once invoked, about $0.0004 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Related

Other skills, from other repositories

systematic-debugging

Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.

obra/superpowers · 21 tokens

chat-pet-sprite-creation

Use when creating or changing VS Code chat pet sprite art, sprite sheets, state animations, eye treatments, Stable/Insiders variants, or pet transitions under src/vs/workbench/contrib/chat/browser/widget/media/chatPet.

microsoft/vscode · 53 tokens

cpu-profile-analysis

Analyze V8/Chrome CPU profiles (.cpuprofile) and DevTools trace files (Trace-.json). Use when: profiling performance, investigating slow functions, comparing code paths, finding bottlenecks, analyzing timeToRequest, understanding call trees from sampling profiler data, analyzing layout/paint/rendering, investigating…

microsoft/vscode · 71 tokens

babysit-pr

Babysit a GitHub pull request after creation by continuously polling review comments, CI checks/workflow runs, and mergeability state until the PR is merged/closed or user help is required. Diagnose failures, retry likely flaky failures up to 3 times, auto-fix/push branch-related issues when appropriate, and keep…

openai/codex · 114 tokens

imagegen

Generate or edit raster images when the task benefits from AI-created bitmap visuals such as photos, illustrations, textures, sprites, mockups, or transparent-background cutouts. Use when Codex should create a brand-new image, transform an existing image, or derive visual variants from references, and the output…

openai/codex · 113 tokens

agent-host-chat-contributions

Build and review cross-cutting agent-host chat behavior through lifecycle contributions. Use when adding turn lifecycle side effects, prompt or context injection, restored-history transformation, protocol-action observation, or when reviewing changes that add code to AgentSideEffects or AgentService.

microsoft/vscode · 56 tokens