novotnyllc/roundhouse/fleet-inventory
Claude Code and agent skill by novotnyllc/roundhouse
- community
Checking that the SKILL.md file is still in the repository.
Inventory and compare a configured fleet of macOS, Linux, WSL, and Windows machines. Use for packages, agent runtimes, allowlisted Claude/Codex settings, plugins, standalone skills and provenance, auth-file or native session status, projects, startup tasks, SHA-256 comparisons, install-time inference, or human and JSONL status reports.
Install
npx vskill@latest install novotnyllc/roundhouse/fleet-inventorySKILL.md
Fleet Inventory
Set SKILL_DIR to the absolute directory containing this loaded SKILL.md and CLI="$SKILL_DIR/../../scripts/roundhouse"; the shell working directory is not the skill directory. Start with "$CLI" validate-config, resolve requested host names or groups from the config, then collect only the sections needed. Default to all configured hosts and a human report; preserve the JSONL snapshot when another agent will consume it. In human reports, group machine entries sharing a physical_host under one hardware heading.
For local and ssh targets, run collect --target HOST --section SECTION. Exit 2 means a usable partial snapshot; show its errors instead of discarding valid records. SSH collection uses the exact installed executor and returns executor_update_required when its version or hashes are stale; inventory does not update the plugin implicitly. Its connection and keepalive waits are bounded; mutating skills additionally require the target's configured native hostname/user to match. Use validate, render, and compare rather than parsing the records ad hoc.
Inventory is private but not cosmetically redacted: show operational values the owner requested. Never include credential contents, tokens, environment values, or authenticated Git URL credentials. Credential records may include path, owner, mode, size, mtime, SHA-256, strategy, and native health. A Codex plugin cache directory birth time is only inferred_installed_at, never an authoritative install date.
For configured JSON/TOML agent settings, show the allowlisted observed and desired values plus in_sync; never dump the surrounding config file. Treat Claude CLI and Desktop Code as sharing supported settings but separate login state. The invoking agent must check and report Codex Desktop Remote enablement manually because it has no documented persistent config key or inventory record.
For codex-remote-control with a wsl_interop_via sibling, "$CLI" collect --target HOST collects natively through the WSL interop lane from any harness; executor_update_required means the installed Windows Roundhouse does not match this controller. Otherwise, or when WSL is unreachable, Codex must read and follow "$SKILL_DIR/../../references/codex-remote-control.md", which defers the task's model and effort choice to railyard:model-routing, and Claude must report that transport as unsupported. Never substitute WSL-side execution for native Windows evidence.
Protected broker readiness
For an enrolled or proposed protected target, run "$CLI" privilege-status HOST SNAPSHOT and report transport, node identity, broker, adapter mechanism, protected machine-package, and bounded profile readiness separately. The shared protected workflow vocabulary for both Codex and Claude is prepare-privilege-identity, prepare-privilege-enrollment, verify-privilege-plan, submit-privilege-plan, lookup-privilege-result, preview-privilege-upgrade, and preview-privilege-revocation. This inventory skill invokes none of the mutating commands.
For macOS, report root-broker enrollment, default-disabled/active state, and only macos.install-signed-pkg.v1 and macos.apply-system-setting.v1 when advertised. Root Homebrew, arbitrary sudo, installer scripts, and arbitrary plist paths are unsupported; owner-local interactive elevation is separate from SSH.
Preserve needs_enrollment, drifted, transport_unavailable, unsupported_context, unsupported_security_boundary, partial, and stale exactly. Show the public audit identity: fleet CA fingerprint and generation, originating node ID, node-key fingerprint, certificate serial and validity, and pinned host-key fingerprint. Never print private-key material. Never ask for or relay a sudo or Administrator password. Enrollment, upgrade, policy activation, and revocation stop at their local human password/UAC boundary; no WSL, shell, Codex-task, or interactive SSH fallback may substitute for the configured protected adapter.
Conclude with observed drift, unavailable evidence, and exact next actions. Do not mutate anything from this skill.
Security
- Security scanPassed, 90 of 100Pattern checks on the skill's files.
- Model reviewPassed
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingplugins/roundhouse/skills/fleet-inventory/SKILL.md
Popularity
- 7-day trend score
- 0
Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/novotnyllc/roundhouse/fleet-inventory)