Agent skill

Claudexor

by razzant in razzant/claudexor

Use the local Claudexor control plane for harness-agnostic coding work across Claude Code, Codex, Cursor, and OpenCode.

MITAuto-check passedAgent Workflows

Install Claudexor

skills CLI
$ npx skills add razzant/claudexor --skill claudexor -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install razzant/claudexor claudexor --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/razzant/claudexor.git skills-src && mkdir -p .claude/skills && cp -r skills-src/plugins/claude/skills/claudexor .claude/skills/claudexor && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
claudexor
GitHub stars
498
Token cost
~1.6k tokens
SKILL.md length
790 words
Files
1
Skills in repo
1
Repo updated
First seen
Licence
MIT

At a glance

Use the local Claudexor control plane for harness-agnostic coding work across Claude Code, Codex, Cursor, and OpenCode.

  • Works in 6 steps: Call claudexor_status to inspect the… → Call claudexor_accounts before choosing… → Call claudexor_capabilities when the… → …
  • A task benefits from route-aware harness and account selection
  • SKILL.md covers Preconditions, Choose a route and Safety
  • Calls npm

What it does

Claudexor is an agent skill from razzant/claudexor. Use the local Claudexor control plane for harness-agnostic coding work across Claude Code, Codex, Cursor, and OpenCode. Use when a task benefits from route-aware harness and account selection, quota-aware account rotation, shared thread context, read-only planning or research, best-of-N execution, or cross-harness review.

Its SKILL.md is about 1.6k tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts.

It sits in Agent Workflows. It works with Model Context Protocol and macOS. The repository describes itself as: Multi-harness control plane for Claude Code, Codex, Cursor, and OpenCode: quota-aware rotation across multiple Claude/Codex subscriptions, shared thread context, and cross-model… The licence is MIT.

When your agent uses it

  • A task benefits from route-aware harness and account selection
  • Quota-aware account rotation
  • Shared thread context
  • Read-only planning

Example prompts

  • “/claudexor”

Requirements

  • Node.js

Workflow steps

6 steps, taken from the first numbered list in SKILL.md.

  1. Call claudexor_status to inspect the aggregate/default-store projection,
  2. Call claudexor_accounts before choosing an account or reviewer identity.
  3. Call claudexor_capabilities when the task depends on the current modes,
  4. Use claudexor_ask for read-only answers and bounded research, or
  5. Only for explicit implementation intent, use claudexor_run,
  6. For durable multi-turn work, call claudexor_thread_create once and enqueue

What it can do on your machine

Read from SKILL.md and the folder at commit 499e347. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Shell commands in SKILL.md call:

    • npm

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md. Its commands use npm, which can reach the network depending on how they are called.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Claudexor loads about 1.6k tokens when it runs. Until then it costs about 83 tokens; SKILL.md has 790 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~83
When it runs · the whole SKILL.md, loaded when a task matches
~1.6k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from razzant/claudexor at commit 499e347, republished under its MIT licence (© razzant). 790 words, ~1,566 tokens.

Download SKILL.mdSave it as .claude/skills/claudexor/SKILL.md (or your agent's skills folder).
name
claudexor
description
Use the local Claudexor control plane for harness-agnostic coding work across Claude Code, Codex, Cursor, and OpenCode. Use when a task benefits from route-aware harness and account selection, quota-aware account rotation, shared thread context, read-only planning or research, best-of-N execution, or cross-harness review.

Claudexor

Use the MCP tools supplied by this plugin. They connect to the user's local Claudexor installation; they do not provide accounts or remote model access.

Preconditions

  • Support macOS and Linux. Do not claim Windows support.
  • Require Node.js 20.19 or newer and a preinstalled claudexor command. If the MCP server is missing, ask the user to run npm install -g claudexor, then restart the host. Do not download a runtime implicitly.
  • If a generated Claudexor host integration already owns the current host's setup, leave it in place. Do not repair, replace, or uninstall it automatically.

Choose a route

Start read-only unless the user explicitly asked to change or create files.

  1. Call claudexor_status to inspect the aggregate/default-store projection, but do not use it as a universal admission veto. An unpinned request may be admitted by the canonical account pool through a ready exact profile even when default doctor is unavailable; only a genuinely profile-less/default fallback requires doctor status ok and the requested intent. Aggregate status neither proves nor vetoes a pooled or explicitly named profile.
  2. Call claudexor_accounts before choosing an account or reviewer identity. It is the read-only Accounts view of registered profiles, readiness, quota freshness, and the daemon's next_up routing projection; the default call combines current quota and registry facts with retained readiness observations. One coalesced cold acquisition, including failure, is followed by passive reads with the original check time; viewing does not renew old evidence. fresh: true requests the atomic snapshot, an expensive explicit refresh to use only when acting on staleness matters. Check the selected row's age: retained available/passed describes its last observation, while actual execution performs its own current admission checks. unknown, not_run, or stale quota means uncertain, not absent. Never substitute aggregate doctor status, another profile's probe, or a host/default login. Never initiate login or OAuth merely because a row is unknown; ask the user for explicit authority. An explicit reviewer credentialProfileId is strict and never falls back, while an omitted id uses the canonical account pool.
  3. Call claudexor_capabilities when the task depends on the current modes, mutability, controls, models, setup transport, or tool surface. The CLI claudexor models --harness <id> is default-route discovery, not a named profile's entitlement; source: none or an unavailable inventory is an honest refusal, not permission to guess. For a named profile, let the strict pinned run/reviewer preflight admit the model and verify observed profile/model telemetry afterward. A declared setupLogin transport of external_terminal means the host must provide the supported client terminal attach path; it is not itself unreadiness and must not be silently treated as in_app.
  4. Use claudexor_ask for read-only answers and bounded research, or claudexor_plan for a read-only implementation plan.
  5. Only for explicit implementation intent, use claudexor_run, claudexor_best_of, or claudexor_create. Pass an absolute repoPath whenever the target repository could be ambiguous. Pass credentialProfileId to pin the run strictly to one account.
  6. For durable multi-turn work, call claudexor_thread_create once and enqueue each follow-up with claudexor_thread_turn. Pass an idempotencyKey and retry an unknown outcome with the same key and body. Follow the returned runId with the same status/result tools as a one-shot run; if only a queued jobId is returned, use claudexor_runs to recover its runId after binding. claudexor_thread_read lists earlier turns without starting work.
Show full SKILL.md (250 more words)Show less

Run tools enqueue work and return a durable handle, not terminal output or proof of completion. Use claudexor_runs, claudexor_inspect, claudexor_run_status, and claudexor_run_result to recover and follow it. Use claudexor_run_interactions to read pending questions and claudexor_answer_interaction only with answers supplied or approved by the user. Use claudexor_run_cancel only when cancellation is requested.

Use claudexor_apply_check to dry-check delivery. MCP does not apply a patch. Only when the user explicitly requests delivery and the server-owned applyEligibility.eligible value is true may the host invoke the ordinary Claudexor CLI apply path under its normal command permission.

For an explicit reviewer panel, pass entries such as {"harness":"cursor","model":"cursor-grok-4.6-high","credentialProfileId":"review-cursor"}. The CLI equivalent is --reviewer-panel-json '<array>'; the legacy compact --reviewer-panel form stays unpinned and should not be extended with an escaping-sensitive profile syntax. After review, verify the observed profile and route proof in the result artifacts.

Use claudexor_journal_recovery for read-only journal inspection, validation, or export. claudexor_quarantine_journal is destructive: invoke it only after the user explicitly requests quarantine and provides the exact fingerprint and confirmation required by its schema.

Safety

  • NEVER paste live credentials into prompts. Ask the user to use claudexor auth login <harness> or claudexor secrets set outside the conversation; do not collect or transform credentials.
  • NEVER auto-answer claudexor decision for a blocked run. Risk acceptance and overrides are human decisions, even when the calling host otherwise permits mutation.
  • Do not infer success, applyability, readiness, or completion from model prose. Use the typed status, result, and applyEligibility projections.
  • Do not claim native vendor sessions move between credential profiles.

© razzant, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

Just SKILL.md in plugins/claude/skills/claudexor of razzant/claudexor.

Open the folder on GitHubat commit 499e347

Compare with similar skills

Claudexor next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Claudexor compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Claudexor this skillrazzant/claudexor498—~1.6kAutomated safety check: PassMIT
Binggo MCPluovicter-collab/bilibinggo500—~1.2kAutomated safety check: PassProprietary
Cortex Mem MCPsopaco/cortex-mem313—~2.8kAutomated safety check: PassMIT
Mps Project ManagementJetBrains/MPS1.7k—~2.2kAutomated safety check: PassApache-2.0
StoryStorybloq/storybloq769—~15kAutomated safety check: PassCustom licence
Premiere Pro MCPhetpatel-11/Adobe_Premiere_Pro_MCP668—~1.3kAutomated safety check: PassMIT

Similar skills

  • Binggo MCP

    luovicter-collab/bilibinggo

    Operates the local Binggo (bilibinggo) Bilibili lottery helper console strictly through binggo MCP tools.

    500 GitHub stars~1.2k tokensUpdated 1 mo ago
    Agent WorkflowsAuto-check passed
  • Cortex Mem MCP

    sopaco/cortex-mem

    Persistent memory enhancement for AI agents. An agent skill from sopaco/cortex-mem.

    313 GitHub stars~2.8k tokensUpdated 2 mo ago
    Agent WorkflowsAuto-check passed
  • Official

    Open an MPS project in a running or freshly started MPS instance when MCP tools fail because no project is open (welcome screen), close an open project with mpsmcpcloseproject, or create a new empty…

    1.7k GitHub stars~2.2k tokensUpdated yesterday
    Agent WorkflowsAuto-check passed
  • Story

    Storybloq/storybloq

    Track tickets, issues, and progress for your project. An agent skill from Storybloq/storybloq.

    769 GitHub stars~15k tokensUpdated 4 days ago
    Agent WorkflowsAuto-check passed
  • Premiere Pro MCP

    hetpatel-11/Adobe_Premiere_Pro_MCP

    Install, verify, troubleshoot, and operate the Adobe Premiere Pro MCP server.

    668 GitHub stars~1.3k tokensUpdated 15 days ago
    Agent WorkflowsAuto-check passed
  • Gearcoleco Romhacking

    drhelius/Gearcoleco

    Hack, modify, and translate ColecoVision and Super Game Module ROMs using the Gearcoleco emulator MCP server.

    142 GitHub stars~3.9k tokensUpdated 3 days ago
    Agent WorkflowsAuto-check passed

Categories

Questions about Claudexor

What does Claudexor do?

Use the local Claudexor control plane for harness-agnostic coding work across Claude Code, Codex, Cursor, and OpenCode. Claudexor is an agent skill from razzant/claudexor. Use the local Claudexor control plane for harness-agnostic coding work across Claude Code, Codex, Cursor, and OpenCode.

When should I use Claudexor?

Claudexor fits situations like: A task benefits from route-aware harness and account selection; quota-aware account rotation; shared thread context; read-only planning.

How do I install Claudexor in Claude Code?

Run `npx skills add razzant/claudexor --skill claudexor -a claude-code`. Or copy the skill folder (plugins/claude/skills/claudexor in razzant/claudexor) into .claude/skills/claudexor in your project. Claude Code loads it when a task matches its description.

How do I install Claudexor in Codex?

Run `npx skills add razzant/claudexor --skill claudexor -a codex`. Or copy the skill folder (plugins/claude/skills/claudexor in razzant/claudexor) into .agents/skills/claudexor in your project. Codex loads it when a task matches its description.

Can I use Claudexor in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add razzant/claudexor --skill claudexor -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/claudexor, .gemini/skills/claudexor, .github/skills/claudexor and .opencode/skills/claudexor in your project.

What does Claudexor need to run?

Going by SKILL.md and its folder, Claudexor needs the command-line tools its instructions call (npm). Our summary lists: Node.js.

Does Claudexor access the network?

SKILL.md contains no URLs. Its commands use npm, which can reach the network depending on how they are called. This is read from the text; nothing was executed.

Is Claudexor safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Claudexor use?

Claudexor is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Claudexor use?

About 1.6k tokens (SKILL.md is roughly 6.3k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Claudexor?

Skills that share tags, products or a category with Claudexor: Binggo MCP (luovicter-collab/bilibinggo, 500 stars), Cortex Mem MCP (sopaco/cortex-mem, 313 stars), Mps Project Management (JetBrains/MPS, 1.7k stars) and Story (Storybloq/storybloq, 769 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Claudexor?

razzant (a GitHub user) maintains it in razzant/claudexor, which has 498 GitHub stars. The repository was last updated on October 9, 2026.

Source: razzant/claudexor on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.