alma-linux-remote-plugin

Stateful Linux remote management for Alma/Pi agents via SSH/SFTP. All remote operations use persistent sessions (no stateless mode), with 5-minute idle timeout, SQLite audit logging, and FastAPI dashboard for audit viewing.

Safety Notice

This listing is imported from skills.sh public index metadata. Review upstream SKILL.md and repository scripts before running.

Copy this and send it to your AI assistant to learn

Install skill "alma-linux-remote-plugin" with this command: npx skills add adfoke/alma-linux-remote-plugin/adfoke-alma-linux-remote-plugin-alma-linux-remote-plugin

Alma Linux Remote Plugin Skill

Purpose

Use this skill to operate remote Linux hosts safely through the plugin tools:

  • list_hosts
  • test_connection
  • run_command
  • upload_file
  • download_file

This project is stateful-first:

  • test_connection now also uses persistent session flow
  • if session exists → reuse
  • if not → auto-create
  • idle timeout defaults to 300 seconds (5 minutes)

Required Execution Order (Agent Policy)

When handling user requests, follow this order unless user explicitly overrides:

  1. list_hosts (discover available targets)
  2. test_connection(host_name) (validate or create session)
  3. run_command(...) / upload_file(...) / download_file(...)
  4. Summarize outputs and risks clearly

Do not assume host aliases that are not in hosts.yaml.


Tool Contracts

1) list_hosts

  • args: {}
  • returns: string[]

2) test_connection

  • args: { "host_name": string, "timeout"?: number }
  • behavior: uses persistent session manager (reuse/create session)
  • returns: success/failure message string

3) run_command

  • args: { "host_name": string, "command": string, "timeout"?: number }
  • returns:
    • command
    • exit_code
    • stdout
    • stderr
    • success

4) upload_file

  • args: { "host_name": string, "local_path": string, "remote_path": string }
  • returns: message string

5) download_file

  • args: { "host_name": string, "remote_path": string, "local_path": string }
  • returns: message string

Session Semantics

  • Session key: per host_name
  • Auto keepalive enabled
  • Idle cleanup thread runs periodically
  • Cleanup threshold: session.idle_timeout_seconds (default 300)

If a command fails due to stale transport, retry once by re-running test_connection then re-run command.


Audit & Observability

Audit is stored in SQLite (no JSONL file mode):

  • default DB: ./logs/audit.db
  • table: audit_logs

FastAPI dashboard:

  • GET / → web page
  • GET /api/logs → query API

Supported query parameters:

  • page
  • page_size
  • host_name
  • operation_type
  • start_time (ISO8601)
  • end_time (ISO8601)

When user asks “查看AI操作日志”, prefer directing to dashboard/API filters first.


Safety Rules

  • Never echo secrets from .env (key passphrase or other secrets)
  • Never fabricate command output
  • Command policy is configurable:
    • policy.default_mode = blocklist (default)
    • policy.default_mode = strict_allowlist
    • policy.host_overrides.<host> for per-host policy
  • If command is blocked, return clear manual steps; do not attempt bypass
  • Always show host + command in response summary

Recommended Response Style

For each remote action, report:

  1. Target host
  2. Command/file action
  3. Exit result (success/failure)
  4. Key stdout/stderr summary
  5. Next suggested step

Examples

Health check

  1. list_hosts
  2. test_connection("prod-web-1")
  3. run_command("prod-web-1", "uptime && df -h")

Upload and verify

  1. test_connection("prod-web-1")
  2. upload_file("prod-web-1", "./deploy.sh", "/tmp/deploy.sh")
  3. run_command("prod-web-1", "chmod +x /tmp/deploy.sh && /tmp/deploy.sh --check")

Audit query (API)

  • /api/logs?page=1&page_size=50&host_name=prod-web-1&start_time=2026-03-03T00:00:00Z&end_time=2026-03-03T23:59:59Z

Source Transparency

This detail page is rendered from real SKILL.md content. Trust labels are metadata-based hints, not a safety guarantee.

Related Skills

Related by shared tags or category signals.

Security

skillguard-hardened

Security guard for OpenClaw skills, developed and maintained by rose北港(小红帽 / 猫猫帽帽). Audits installed or incoming skills with local rules plus Zenmux AI intent review, then recommends pass, warn, block, or quarantine.

Archived SourceRecently Updated
Security

api-contract-auditor

审查 API 文档、示例和字段定义是否一致,输出 breaking change 风险。;use for api, contract, audit workflows;do not use for 直接改线上接口, 替代契约测试平台.

Archived SourceRecently Updated
Security

ai-workflow-red-team-lite

对 AI 自动化流程做轻量红队演练,聚焦误用路径、边界失败和数据泄露风险。;use for red-team, ai, workflow workflows;do not use for 输出可直接滥用的攻击脚本, 帮助破坏系统.

Archived SourceRecently Updated