ManageLM
Natural-language Linux and Windows server management for Claude.
Community: Submitted by a user or imported; check the owner before granting accessOnlineNo sign-inGlobalFreeRead-only
What it can do
- Get Agent Info: Get detailed info for a single agent: OS, version, health metrics (CPU/memory/disk), LLM status, assigned skills with their operations, and recent task history — all in one call. Use t
- Get Agent Skills: Get the skill state for a specific agent: every account skill annotated as assigned (direct or via group) or not yet assigned. Use this to understand what an agent can do and what ad
- List Available Skills: List skills from the built-in catalog that have NOT been imported into your account yet. Use this to discover additional capabilities you could enable by importing them from the
What data it sees
Do you need an account
No: the server works without sign-in
Natural-language Linux and Windows server management for Claude. Agents on each host execute tasks and run fleet-wide inventory, security-audit and access-review scans.
Server tool list (33)
Raw names from tools/list. Only developers need these.
| get_agent_info | Get detailed info for a single agent: OS, version, health metrics (CPU/memory/disk), LLM status, assigned skills with their operations, and recent task history — all in one call. Use this to understand an agent's full capabilities and state before running tasks. |
| get_agent_skills | Get the skill state for a specific agent: every account skill annotated as assigned (direct or via group) or not yet assigned. Use this to understand what an agent can do and what additional skills could be enabled. |
| list_available_skills | List skills from the built-in catalog that have NOT been imported into your account yet. Use this to discover additional capabilities you could enable by importing them from the portal's skill catalog. |
| get_account_info | Get your account plan, usage limits, current consumption, and the catalog of groups and sites in the account (with member counts). Use this to check remaining capacity before running large batch operations, and to discover valid `group=` / `site=` filter values for tools like search_agents. |
| list_team_members | List ManageLM users in your account with their role, permissions, and registered SSH public keys. Use cases: - "Who has admin access?" — check roles and permissions - "Which team members haven't registered an SSH key?" — find users with no keys - "Get Charly's SSH public key" — retrieve key details |
| search_agents | List, search and filter agents (servers, hosts, machines) by status, group, site, health metrics, or free text. Returns every accessible agent when called without filters. Each row includes status, OS, IP addresses, health metrics, LLM readiness, groups, site, tags, last-seen time. Filters: query (free-text on hostname, OS, IP, tags), status (online/offline/approved/pending_approval), group, site, cpu_above, memory_above, disk_above. Call this first to discover agent targets before dispatching tasks. |
| search_inventory | Search system inventory across all your agents. Find which servers run a specific service, have a package installed, or run containers — without dispatching any commands. Queries the latest inventory reports stored in the portal. |
| search_security | Search security issues across all your agents — a unified view of three sources, each row tagged with its `source`: (1) "audit" — security audit findings, (2) "pentest" — pentest report findings, (3) "threat" — real-time threat detection alerts (service/session kind). Use this for "what security issues do I have", "which servers have critical issues", "any open threats", "SSH misconfigurations", "what fired in the last 24h", etc. Time-range filtering via since/until applies to each source's timestamp. |
| search_activity | Search user activity across all your agents — successful logins, failed logins, sudo/elevated commands, and file changes. Use this tool whenever the user asks about login activity, login counts, who logged in, failed authentication, sudo usage, or user activity on servers. Queries the latest activity audit reports stored in the portal — without dispatching any commands. Works with both Linux and Windows agents. Supports time-range filtering via since/until (ISO-8601) for queries like "who logged in yesterday". |
| search_ssh_keys | Search SSH keys and access credentials across your infrastructure. Combines two data sources: 1. Server-side: authorized_keys (Linux) or local user accounts (Windows) found by access scans 2. ManageLM profiles: registered SSH public keys (team members' keys with full public key content) Use cases: - "Who has SSH access to pocmail?" — shows deployed keys with identity mapping - "Get Charly's SSH key" — returns the full public key from their ManageLM profile - "List unknown SSH keys" — keys on servers not matched to any team member - "Does Charly have SSH access to pocmail?" — checks both profile and deployed keys When GRANTING SSH access for a team member, this tool provides the full public key needed by the users skill. Call this tool first, then pass the key to the users skill. IMPORTANT: When looking up a person's keys, always use the "user" parameter (matches ManageLM profile name/email), never "query". |
| search_sudo_rules | Search privileged access across your infrastructure. On Linux: queries sudoers rules from access scans. On Windows: queries Administrators and privileged group membership. Use cases: - "Who has sudo on production servers?" — use group filter - "Show all NOPASSWD sudo rules" — use nopasswd_only - "What sudo does toto have on pocmail?" — use query for system username - "What are Charly's sudo privileges?" — use user param (matches via SSH key identity) - "Who are the administrators on the Windows servers?" — use group filter To GRANT or REVOKE sudo, use the users skill after checking current state with this tool. |
| search_certificates | Search the x509/TLS certificates discovered on your servers by certificate scans — every cert sitting in system stores, web-server config dirs, Let's Encrypt live dirs, etc., whatever issued it. This is the source of truth for what is actually deployed on hosts. Use it to audit real on-disk certs: find expiring, expired, self-signed, or weak-key/weak-signature certificates, or certs NOT managed by ManageLM. For the certificates ManageLM itself issues and renews (its Internal CA + Let's Encrypt), use search_pki instead. Run run_certificate_scan first if no scan has been performed yet. Each row includes: hostname, subject, issuer, path, not_after, days_left, key_type, key_bits, sig_alg, is_ca, self_signed, managed, expired, expiring_soon, weak_key, weak_sig. |
| run_security_audit | Run a security audit (also called "security report") on one or more agents. Checks SSH config, firewall rules, open ports, user accounts, file permissions, TLS certificates, Docker, kernel settings, and more. Waits for completion and returns structured findings with severity, category, title, explanation, and remediation steps. Requires Reports Admin permission. Use search_security to query existing reports without re-running. |
| run_inventory_scan | Run a system inventory scan (also called "inventory report") on one or more agents. Discovers running services, installed packages, containers, network interfaces, storage mounts, hardware info, scheduled tasks, and more. Waits for completion and returns structured inventory items with category, name, status, version, and details. Requires Reports Admin permission. Use search_inventory to query existing reports without re-running. |
| run_access_scan | Run an access scan (also called "access report") on one or more agents. On Linux: discovers authorized_keys and sudoers rules for every system user. On Windows: discovers local user accounts, admin group membership, and RDP access. Returns access data with identity mapping via ManageLM profiles. Triggers: "run access scan", "start access report", "scan SSH keys", "scan sudo access", "initiate access report". Requires Reports Admin permission. Use search_ssh_keys or search_sudo_rules to query existing results without re-running. |
| run_certificate_scan | Run a certificate scan on one or more agents to discover the x509/TLS certificates present on disk — walks the host's certificate stores, web-server config dirs, and Let's Encrypt live dirs (metadata only — private key material is never collected). Waits for completion and returns the discovered certificates with subject, issuer, path, expiry, key type/bits, and flags for self-signed / expiring / expired / weak. Requires Reports Admin permission. Use search_certificates to query existing results without re-running. This scans certs ON the servers — it does NOT issue or manage certificates (that is PKI; see search_pki). |
| run_activity_scan | Run a fresh activity scan on one or more agents. Dispatches a live collection command — use search_activity instead to query existing reports. On Linux: collects from standard logs (no auditd required). On Windows: collects from Windows Event Viewer. Returns structured activity data (logins, failed auth, privileged commands, file changes). Requires Reports Admin permission. |
| get_task_status | Get the status of a previously submitted task. |
| get_task_history | Get recent task history for an agent (default 20, max 200). Each row shows skill, operation, status, summary, and any pending question. Supports time-range filtering via since/until (ISO-8601). Use get_task_status with a specific task_id for the full request/response payload, or get_task_changes to see file modifications. |
| get_task_changes | Get the file changes made by a specific task. Shows which files were modified in /etc/ and other tracked directories. Returns the list of changed files and a summary. Use get_task_history to find task IDs. |
| answer_task | Answer a question from an interactive task that returned needs_input status. The agent paused because it needs information it cannot determine on its own (e.g. domain name, password, license key). Ask the user for the answer, then call this tool. May return another question if the setup requires multiple inputs. |
| revert_task | Revert the file changes made by a specific task. Fetches the diff from the agent's local git repo and applies a reverse patch to restore the previous state. The revert itself is tracked as a new task. Only works if the agent is online and the changeset is still within the 30-day retention window — call get_task_changes first to confirm the task is recent enough and to preview what will be undone. |
| search_pki | List, search and filter the TLS/SSL certificates ManageLM ITSELF ISSUES AND MANAGES via its PKI (Internal CA + Let's Encrypt) — by status, source, agent, site, or free text. Returns every managed certificate when called without filters. Each row includes status, expiry, days_left, source (Internal CA or Let's Encrypt), target agent, SANs. Filters: query (free-text on common name, agent, SAN), status (active/revoked/expired/failed/pending), source (local_ca/letsencrypt), agent, site. Use this to find expiring managed certs, audit issued-certificate health, or locate a managed cert on a specific server. This covers only certificates ManageLM issued. For every cert actually found on the servers (managed or not, any issuer), use search_certificates. |
| search_monitors | List, search and filter monitors, health checks, or service probes (network checks AND resource checks) by status, type, agent, site, or free text. Returns every monitor when called without filters. Each row includes status (up/down/degraded/pending), measured value, value unit, agent, check type, schedule. The `value` field is response time in ms for network monitors (http/tcp/dns/udp), or % usage for cpu/memory/filesystem, or instance count for process monitors — the `unit` field disambiguates ("ms" / "%" / "count"). Resource monitors map down=critical, degraded=warning. Filters: query (free-text on monitor name, agent, type), status (up/down/degraded/pending), slug (e.g. "website", "mysql", "smtp", "cpu", "memory", "filesystem", "process"), agent, site. Use this to find down/critical monitors or audit infrastructure health. |
| search_backups | List, search and filter filesystem backups, backup jobs, or snapshot schedules by status, agent, site, or free text. Returns every backup when called without filters. Each row includes status (ok/failed/running/pending), source path, schedule, keep count, snapshot count, last run time, target agent. Filters: query (free-text on backup name, source path, agent), status (ok/failed/running/pending), agent, site. Use this to find failing backups, audit backup health, or locate a backup by name or source path. |
| send_email | Send an email to yourself (the authenticated MCP user). Use this to deliver reports, summaries, or notifications. The body should be plain text — newlines and paragraphs are preserved. The email is wrapped in the ManageLM template. |
| base | Core read-only utilities for file reading, search, system info, resource usage, and network diagnostics — always available across all skills Operations: - inspect: linux: Read files, list directories, search content, view file metadata; windows: Read files, list directories, search content, view registry keys and file metadata - system_info: linux: Check OS, kernel, hostname, uptime, hardware, users, and login history; windows: Check OS, hostname, uptime, hardware, and logged-in users - resources: linux: Monitor disk, memory, CPU, processes, and I/O statistics; windows: Monitor disk, memory, CPU, processes, and performance counters - network: linux: Check interfaces, connections, DNS, ping, traceroute, and fetch URLs; windows: Check adapters, connections, DNS, ping, traceroute, and fetch URLs The agent adapts to the target server OS (Linux/Windows) automatically. Use plain-language instructions — do not specify OS-specific commands. |
| files | Create, copy, move, delete files and folders, manage permissions, extended attributes, filesystem attributes, and compression Operations: - read: Read the contents of a file - write: Write content to a file (creates or overwrites) - edit: Edit a file — replace text, append, insert, or delete lines - list: List directory contents with details - permissions: Change file or directory permissions and ownership - compress: Compress files or directories into an archive - extract: Extract an archive - find: Search for files by name, size, date, type, or content - disk_usage: Analyze disk usage for a path - upload: Upload a file to the server (binary-safe, base64-encoded). Use this when the user wants to transfer a file to the server — e.g. deploy a website zip, upload a config file, etc. - download: Download a file from the server (binary-safe, base64-encoded). Use this when the user wants to retrieve a file — e.g. download a config, get a log file, etc. The agent adapts to the target server OS (Linux/Windows) automatically. Use plain-language instructions — do not specify OS-specific commands. File transfer: To upload a file, provide content (plain text) or content_base64 (binary) + path. This performs a direct file write — no LLM involved. To download, set instruction to "download <path>". IMPORTANT: When the user asks to upload, transfer, or write a file to a server, ALWAYS use the content + path parameters instead of putting file content in the instruction. |
| logs | View, search, filter, and manage system logs, log rotation, and web log analysis Operations: - system: View system journal logs, optionally filtered by unit, priority, or time range - application: View application log files - search: Search logs for a pattern or keyword - errors: Show recent errors and warnings from system or application logs - analyze: Analyze logs — summarize patterns, count occurrences, identify anomalies The agent adapts to the target server OS (Linux/Windows) automatically. Use plain-language instructions — do not specify OS-specific commands. |
| network | Manage network adapters, IP addresses, routing, bridges, and traffic shaping (firewall lives in the firewall skill; DNS server management in the dns skill) Operations: - status: Show network status: interfaces, IPs, routes, DNS, and listening ports - connectivity: Test connectivity to a host — ping, DNS lookup, traceroute, or port check - ports: List open/listening ports and the processes using them - connections: List active network connections - configure: Configure network settings — IP addresses, routes, DNS, hostname resolution The agent adapts to the target server OS (Linux/Windows) automatically. Use plain-language instructions — do not specify OS-specific commands. |
| packages | Install, update, and remove software packages and system updates Operations: - install: Install one or more packages - remove: Remove one or more packages - update: Update package lists and optionally upgrade all packages - search: Search for packages by name or description - info: Show detailed information about a package - list_installed: List installed packages, optionally filtered The agent adapts to the target server OS (Linux/Windows) automatically. Use plain-language instructions — do not specify OS-specific commands. |
| security | Manage security hardening — fail2ban, SSH hardening, SELinux/AppArmor, auditd, ClamAV antivirus, AIDE integrity, rkhunter, chkrootkit, Lynis, PAM, and kernel sysctls (firewall rules live in the firewall skill) Operations: - fail2ban: Manage fail2ban — status, jails, banned IPs, configuration - ssh_config: View or harden SSH server configuration - ssl: Manage SSL/TLS certificates — list, generate, renew, inspect - selinux: Manage SELinux or AppArmor - auth_logs: Review authentication and security logs — failed logins, sudo usage, suspicious activity The agent adapts to the target server OS (Linux/Windows) automatically. Use plain-language instructions — do not specify OS-specific commands. |
| users | Manage user accounts, groups, passwords, and file permissions Operations: - list_users: List user accounts on the system - add_user: Create a new user account - remove_user: Remove a user account - modify_user: Modify user account properties — UID, GID, groups, shell, home directory, comment/GECOS, lock/unlock, expiry - ssh_key: Manage SSH authorized keys for a user - sudo: Manage sudo access for a user or group - groups: Manage groups — list, create, delete, or modify groups The agent adapts to the target server OS (Linux/Windows) automatically. Use plain-language instructions — do not specify OS-specific commands. |