Holmes-Kit Companion Guide

    1

    A companion for installing and learning Holmes-Kit with Claude Code CLI, Codex CLI or AGY CLI, using Holmes Lab's current installation guide and basic/advanced learning routes.

    Free

    Security scanned

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    About this skill

    Install first, then learn and apply

    Start before installation or continue with an existing setup. The guide checks the actual user-selected workspace and connects the agent to Holmes Lab's AI installation reference: https://holmes-lab.web.app/agents/holmes-kit/install.md .

    It explains guide-basis versus latest-at-install version policies and separate user folder, project-local and global npm locations, then waits for the human's independent choices and necessary change approvals. Existing choices and intentional pinned/development runtimes are preserved.

    Continue with basic and advanced workflows

    After installation, verify the actual CLI, applied configuration and reconnected MCP binding. Continue with the first verified change or one of eight advanced workflows: adoption, requirements, change impact, verification, bounded autonomy, orchestration, semantics and operations. Report installed/configured/restart-required/binding-verified/live-hook evidence separately.

    Compatibility and limits

    Supported Holmes-Kit client targets: Claude Code CLI, Codex CLI and AGY CLI. Marketplace host badges are not runtime compatibility evidence. Exact OS/client behavior follows the selected release's documentation and actual observation. Companion package version 1.1.0 is independent of the Holmes-Kit 0.40.0 documentation baseline and the human-selected installation version.

    This package contains instructions, three Markdown documents and an unchanged license notice, not the runtime, hooks or MCP server. No bundled executable, cloud key handling, automatic installation or automatic autonomy. If web access is unavailable, disclose that limitation and use the bundled decision guide; obtain current official command documentation before installing. No public source repository availability is asserted.

    Installation guide: https://holmes-lab.web.app/agents/holmes-kit/install.md Korean learning hub: https://holmes-lab.web.app/ko/holmes-kit/ English learning hub: https://holmes-lab.web.app/en/holmes-kit/ Official npm distribution: https://www.npmjs.com/package/@holmes-lab/holmes-kit

    Instruction-only companion; loading it installs no runtime and enables no hooks or sandbox. Installation guidance can begin before Holmes MCP exists; governed development needs an actually installed runtime and connected server. Current command recipes require authorized official-document access; when unavailable, the skill explains choices and stops before unverified installation. Documentation review basis: Holmes-Kit 0.40.0; selected-release compatibility and native trust must be checked. Package structure, public references and isolated behavior scenarios were checked, but a live Agensi delivery/runtime installation test and hardware compatibility test have not been performed.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before any payment.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Free forever

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 12 days ago

    • Free with an Agensi account

    Requires

    Claude Code or Codex CLI or AGY CLI
    Terminal access
    Node.js
    npm
    Git
    Holmes-Kit runtime
    Holmes Kit MCP
    fallback.

    Permissions

    Read Files
    Terminal / Shell
    Write Files
    Network Access

    Terminal and file access are used only for the human-authorized workspace and approved installation/configuration or development. Network access is for public Holmes Lab and official package/release documentation, plus npm installation only after the human selects the version policy/location and approves the relevant changes. No repository or credential upload, environment-variable inspection, unrelated browsing, cloud embedding or automatic autonomy. Browser automation is not required when an authorized HTTPS/document reader exists. Without web access, disclose the limitation and stop before unverified installation. These marketplace declarations do not grant permissions on the user host.

    Listed13 days ago

    What's inside

    Frequently Asked Questions

    It connects installation and learning to Holmes Lab: https://holmes-lab.web.app/agents/holmes-kit/install.md for setup, then the bilingual basics and eight advanced workflow guides. It also retains workspace checks, requirement lineage, operator approvals and evidence reporting. Supported Holmes-Kit targets are Claude Code CLI, Codex CLI and AGY CLI; loading the skill alone enables no hooks or sandbox.

    A supported local CLI host with terminal and repository access and the intended workspace. Holmes MCP is required for the connected development workflow, not for the initial installation decision. Inspect Node/npm/Git and the selected release requirements; propose missing prerequisites instead of silently installing them. Current installation recipes require authorized access to the official web documentation.

    When the connected server returns a pending request, the agent uses approval_ask with its exact request ID. The operator decides. The skill does not issue grants or bypass refusals.

    It contains instructions, two supporting references and a license notice, not the runtime. It can guide your agent before Holmes-Kit exists, but installation proceeds only after the human selects a version policy and location and approves the relevant changes. Loading the skill alone installs nothing. Agensi does not host the Holmes-Kit server or activate local hooks.

    It guides the agent to inspect workspace_preflight, build information and existing diagnostics. It reports missing or stale evidence rather than claiming a successful repair.

    The companion has its own version. Holmes-Kit 0.40.0 is the documentation review basis, not a fixed installation requirement. The human independently chooses guide-basis or latest-at-install policy and an installation location, then approves the exact proposed versions and necessary changes. Existing pinned/development runtimes are preserved. Unavailable release documentation or changed proposals stop installation rather than trigger guessed commands or silent substitutions.

    AI installation: https://holmes-lab.web.app/agents/holmes-kit/install.md . Korean basics: https://holmes-lab.web.app/ko/holmes-kit/basics/ . English basics: https://holmes-lab.web.app/en/holmes-kit/basics/ . Their learning hubs link eight advanced workflows. The official runtime is distributed separately at https://www.npmjs.com/package/@holmes-lab/holmes-kit . Public source repository access is not asserted.

    New in Code Quality & Review

    Free