Agent skill

Octomind

by Muvon in Muvon/octomind

Delegate coding to the Octomind CLI — a token-efficient, model-agnostic coding agent.

Apache-2.0Auto-check passedAgent Workflows

Install Octomind

skills CLI
$ npx skills add Muvon/octomind --skill octomind -a claude-code

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

GitHub CLI
$ gh skill install Muvon/octomind octomind --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/Muvon/octomind.git skills-src && mkdir -p .claude/skills && cp -r skills-src/skills/octomind .claude/skills/octomind && 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
octomind
GitHub stars
159
Token cost
~1.6k tokens
SKILL.md length
554 words
Files
1
Skills in repo
1
Repo updated
First seen
Licence
Apache-2.0

At a glance

Delegate coding to the Octomind CLI — a token-efficient, model-agnostic coding agent.

  • Works in 5 steps: Verify readiness:… → For bounded tasks, pipe the task into… → For iterative tasks, start a named… → …
  • The user asks to use Octomind
  • SKILL.md covers When to Use, Prerequisites, One-Shot Tasks and Iterative Work (Daemon + send), plus 7 more sections
  • Calls brew and cargo; needs OPENROUTER_API_KEY and OCTOHUB_API_KEY

What it does

Octomind is an agent skill from Muvon/octomind. Delegate coding to the Octomind CLI — a token-efficient, model-agnostic coding agent. Use when the user asks to use Octomind, or to hand off features, fixes, refactors, reviews, or long-running coding work to an external agent.

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. Compatibility notes: Requires the octomind CLI (Linux, macOS, Windows) and model access via octomind login or a provider API key

It sits in Agent Workflows, covering Refactoring. It works with Model Context Protocol and Rust. The repository describes itself as: Open-source AI coding agent and agent runtime: one binary, any model, MCP-native. Runs in terminal, CI, or as a daemon. The licence is Apache-2.0.

When your agent uses it

  • The user asks to use Octomind
  • Hand off features
  • Long-running coding work to an external agent

Example prompts

  • “/octomind”

Requirements

  • A credential in OPENROUTER_API_KEY
  • A credential in OCTOHUB_API_KEY
  • Compatibility (from SKILL.md): Requires the octomind CLI (Linux, macOS, Windows) and model access via `octomind login` or a provider API key

Workflow steps

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

  1. Verify readiness: terminal(command="octomind --version").
  2. For bounded tasks, pipe the task into octomind run developer:general --format plain.
  3. For iterative tasks, start a named --daemon session in the background and use octomind send.
  4. Monitor long tasks with process(action="poll"|"log").
  5. Summarize file changes, test results, and remaining risks back to the user.

What it can do on your machine

Read from SKILL.md and the folder at commit 0a01e47. 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:

    • brew
    • cargo

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

  • Network

    No URLs in SKILL.md.

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

  • Credentials

    Names these keys or tokens, usually read from environment variables:

    • OPENROUTER_API_KEY
    • OCTOHUB_API_KEY

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

  • Compatibility

    Requires the octomind CLI (Linux, macOS, Windows) and model access via `octomind login` or a provider API key

    From compatibility in the SKILL.md frontmatter.

Context cost

Octomind loads about 1.6k tokens when it runs. Until then it costs about 59 tokens; SKILL.md has 554 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~59
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 Muvon/octomind at commit 0a01e47, republished under its Apache-2.0 licence (© Muvon). 554 words, ~1,570 tokens.

Download SKILL.mdSave it as .claude/skills/octomind/SKILL.md (or your agent's skills folder).
name
octomind
description
Delegate coding to the Octomind CLI — a token-efficient, model-agnostic coding agent. Use when the user asks to use Octomind, or to hand off features, fixes, refactors, reviews, or long-running coding work to an external agent.
compatibility
Requires the octomind CLI (Linux, macOS, Windows) and model access via `octomind login` or a provider API key
license
Apache-2.0
metadata.author
Muvon
metadata.version
1.0.0

Octomind CLI

Use Octomind as an autonomous coding worker orchestrated by Hermes terminal/process tools. Octomind is an open-source, model-agnostic coding agent in one Rust binary, built for long autonomous runs at low token cost (adaptive cache-aware context compaction, out-of-band planning and loop detection).

When to Use

  • User explicitly asks to use Octomind
  • You want an external coding agent to implement, fix, refactor, or review code
  • Long-running coding sessions where token cost matters
  • Parallel task execution in isolated workdirs/worktrees

Prerequisites

One-Shot Tasks

octomind run takes no message argument — its positional argument is the agent tag. Pipe the task through stdin; piped input runs non-interactively and exits when the work is done:

terminal(command="echo 'Add retry logic to API calls and update tests' | octomind run developer:general --format plain", workdir="~/project")

Machine-readable events instead of text:

terminal(command="echo 'List the TODO items in src/' | octomind run developer:general --format jsonl", workdir="~/project")

Force a specific model:

terminal(command="echo 'Refactor the auth module' | octomind run developer:general --format plain --model openrouter:anthropic/claude-sonnet-4", workdir="~/project")

Restrict file writes to the working directory with --sandbox.

Iterative Work (Daemon + send)

For multi-step work, keep one named session alive and inject follow-ups from separate commands:

terminal(command="echo 'Implement the OAuth refresh flow' | octomind run developer:general --name oauth --daemon --format jsonl", workdir="~/project", background=true)
# Returns session_id

# Monitor progress
process(action="poll", session_id="<id>")
process(action="log", session_id="<id>")

# Send a follow-up to the running session
terminal(command="octomind send --name oauth 'Now add error handling for token expiry'", workdir="~/project")

# Stop
process(action="kill", session_id="<id>")

Resuming Sessions

Sessions persist on disk. --name <name> resumes an existing session of that name; --resume-recent picks the most recent session for the working directory:

terminal(command="echo 'Continue with the remaining tests' | octomind run developer:general --name oauth --format plain", workdir="~/project")

Common Flags

FlagUse
run <tag>Start a specialist: developer:general for coding; other domain:spec tags come from Octomind's tap registry
--format plain / --format jsonlNon-interactive output for piped input
--name <name> / -nCreate or resume a named session
--resume-recentResume the latest session in this directory
--model <provider:model> / -mOverride the model
--daemonKeep the session alive for octomind send
--sandboxLimit filesystem writes to the working directory
--schema <file>Constrain output to a JSON Schema (model must support structured output)

Hosts that speak the Agent Client Protocol (ACP) can run the same specialist as a sub-agent over stdio: octomind acp developer:general.

Show full SKILL.md (235 more words)Show less

Procedure

  1. Verify readiness: terminal(command="octomind --version").
  2. For bounded tasks, pipe the task into octomind run developer:general --format plain.
  3. For iterative tasks, start a named --daemon session in the background and use octomind send.
  4. Monitor long tasks with process(action="poll"|"log").
  5. Summarize file changes, test results, and remaining risks back to the user.

Parallel Work Pattern

Use separate workdirs/worktrees to avoid collisions:

terminal(command="echo 'Fix issue #101 and commit' | octomind run developer:general --format plain", workdir="~/.hermes/cache/scratch/issue-101", background=true)
terminal(command="echo 'Add parser regression tests and commit' | octomind run developer:general --format plain", workdir="~/.hermes/cache/scratch/issue-102", background=true)
process(action="list")

Pitfalls

  • Passing the task as an argument (octomind run 'fix it') treats the text as an agent tag — always pipe the task.
  • --format at a terminal without piped input errors unless combined with --daemon.
  • The first run of a specialist fetches its tap and installs tool dependencies, so it is slower than later runs.
  • A missing OCTOHUB_API_KEY with the default octohub:auto model means octomind login was not run.
  • Avoid sharing one working directory across parallel Octomind sessions.
  • Run inside the project's git repository: the code-search server octocode refuses to start outside one, leaving the agent without semantic search.

Verification

Smoke test:

terminal(command="echo 'Respond with exactly: OCTOMIND_SMOKE_OK' | octomind run developer:general --format plain")

Success criteria:

  • Output includes OCTOMIND_SMOKE_OK
  • Command exits without provider/model errors
  • For code tasks: expected files changed and tests pass

Rules

  1. Prefer piped one-shot runs for automation — they exit on their own.
  2. Use a named daemon session only when iteration is needed.
  3. Always scope Octomind sessions to a single repo/workdir.
  4. For long tasks, provide progress updates from process logs.
  5. Report concrete outcomes (files changed, tests, remaining risks).

© Muvon, Apache-2.0. 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 skills/octomind of Muvon/octomind.

Open the folder on GitHubat commit 0a01e47

Compare with similar skills

Octomind 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.

Octomind compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Octomind this skillMuvon/octomind159—~1.6kAutomated safety check: PassApache-2.0
Refactortermide/termide171—~2.8kAutomated safety check: PassMIT
Rocketmq Rust Issue Generatormxsm/rocketmq-rust1.5k—~1.6kAutomated safety check: PassApache-2.0
Embedded DebuggerAdancurusul/embedded-debugger-mcp202—~1.4kAutomated safety check: PassMIT
Agnixagent-sh/agnix445—~874Automated safety check: PassApache-2.0
Agnixagent-sh/agnix445—~563Automated safety check: PassApache-2.0

Similar skills

  • Refactor

    termide/termide

    Full-workspace code quality analysis and refactoring with validation

    171 GitHub stars~2.8k tokensUpdated today
    DevelopmentAuto-check passed
  • A skill your agent uses when the user asks to create, draft, prepare, or publish a GitHub issue for the rocketmq-rust project — bugs, features, enhancements, refactors, docs, unit tests, CI…

    1.5k GitHub stars~1.6k tokensUpdated today
    Testing & QAAuto-check passed
  • Embedded Debugger

    Adancurusul/embedded-debugger-mcp

    Embedded hardware debugging workflow for probe-rs targets using embedded-debugger-mcp.

    202 GitHub stars~1.4k tokensUpdated 2 mo ago
    Agent WorkflowsAuto-check passed
  • Agnix

    agent-sh/agnix

    A skill your agent uses when user asks to 'lint agent configs', 'validate skills', 'check CLAUDE.md', 'validate hooks', 'lint MCP'.

    445 GitHub stars~874 tokensUpdated today
    Agent WorkflowsAuto-check passed
  • Agnix

    agent-sh/agnix

    A skill your agent uses when user asks to 'lint agent configs', 'validate skills', 'check CLAUDE.md', 'validate hooks', 'lint MCP'.

    445 GitHub stars~563 tokensUpdated today
    Agent WorkflowsAuto-check passed
  • Harness Engineering Lifecycle

    Arenukvern/mcp_flutter

    Design, implement, and integrate generalized validation harnesses across a producer-consumer boundary after a local harness contract exists.

    387 GitHub stars~1.6k tokensUpdated 8 days ago
    Agent WorkflowsAuto-check passed

Questions about Octomind

What does Octomind do?

Delegate coding to the Octomind CLI — a token-efficient, model-agnostic coding agent. Octomind is an agent skill from Muvon/octomind. Delegate coding to the Octomind CLI — a token-efficient, model-agnostic coding agent.

When should I use Octomind?

Octomind fits situations like: the user asks to use Octomind; hand off features; long-running coding work to an external agent.

How do I install Octomind in Claude Code?

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

How do I install Octomind in Codex?

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

Can I use Octomind 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 Muvon/octomind --skill octomind -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/octomind, .gemini/skills/octomind, .github/skills/octomind and .opencode/skills/octomind in your project.

What does Octomind need to run?

Going by SKILL.md and its folder, Octomind needs the command-line tools its instructions call (brew and cargo) and credentials named OPENROUTER_API_KEY and OCTOHUB_API_KEY. Our summary lists: A credential in OPENROUTER_API_KEY; A credential in OCTOHUB_API_KEY. Compatibility (from SKILL.md): Requires the octomind CLI (Linux, macOS, Windows) and model access via `octomind login` or a provider API key.

Does Octomind access the network?

SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.

Is Octomind 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 Octomind use?

Octomind is published under the Apache-2.0 licence (declared in SKILL.md). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Octomind 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 Octomind?

Skills that share tags, products or a category with Octomind: Refactor (termide/termide, 171 stars), Rocketmq Rust Issue Generator (mxsm/rocketmq-rust, 1.5k stars), Embedded Debugger (Adancurusul/embedded-debugger-mcp, 202 stars) and Agnix (agent-sh/agnix, 445 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Octomind?

Muvon (a GitHub organization) maintains it in Muvon/octomind, which has 159 GitHub stars. The repository was last updated on October 10, 2026.

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