Skip to main content
Glama

task_exec_sequence

Execute multiple SSH commands sequentially on a single remote server to automate server management tasks.

Instructions

Exécute plusieurs commandes SSH en séquence sur le même serveur.

Input Schema

TableJSON Schema
NameRequiredDescriptionDefault
aliasYesAlias du serveur cible.
commandsYesListe des commandes à exécuter en séquence (minimum 1).
continueOnErrorNoContinuer même si une commande échoue.
rappelNoDéfinit un rappel en secondes.

Implementation Reference

  • The primary handler function registered for the 'task_exec_sequence' MCP tool. It queues a job of type 'ssh_sequence', logs it to history, initiates execution via the SSH module, waits for completion (with adjusted timeout), and returns the results or background initiation message.
    async (params) => { const job = queue.addJob({ type: 'ssh_sequence', ...params, status: 'pending' }); history.logTask(job); ssh.executeCommandSequence(job.id); const finalJob = await waitForJobCompletion(job.id, config.syncTimeout * params.commands.length); if (finalJob) { return { content: [{ type: "text", text: `Résultat séquence (tâche ${finalJob.id}):\n${JSON.stringify(finalJob, null, 2)}` }] }; } else { return { content: [{ type: "text", text: `Séquence de commandes ${job.id} initiée en arrière-plan.` }] }; } }
  • Zod input schema validating the tool parameters: server alias, array of commands (simple strings or detailed objects with timeout and continueOnError), global continueOnError flag, and optional reminder seconds.
    inputSchema: z.object({ alias: z.string().describe("Alias du serveur cible."), commands: z.array(z.union([ z.string(), z.object({ command: z.string(), timeout: z.number().optional(), continueOnError: z.boolean().optional() }) ])).min(1).describe("Liste des commandes à exécuter en séquence (minimum 1)."), continueOnError: z.boolean().optional().default(false).describe("Continuer même si une commande échoue."), rappel: z.number().optional().describe("Définit un rappel en secondes.") })
  • server.js:580-615 (registration)
    MCP server registration of the 'task_exec_sequence' tool, specifying title, description, input schema, and the inline handler function.
    server.registerTool( "task_exec_sequence", { title: "Exécuter une séquence de commandes (SSH)", description: "Exécute plusieurs commandes SSH en séquence sur le même serveur.", inputSchema: z.object({ alias: z.string().describe("Alias du serveur cible."), commands: z.array(z.union([ z.string(), z.object({ command: z.string(), timeout: z.number().optional(), continueOnError: z.boolean().optional() }) ])).min(1).describe("Liste des commandes à exécuter en séquence (minimum 1)."), continueOnError: z.boolean().optional().default(false).describe("Continuer même si une commande échoue."), rappel: z.number().optional().describe("Définit un rappel en secondes.") }) }, async (params) => { const job = queue.addJob({ type: 'ssh_sequence', ...params, status: 'pending' }); history.logTask(job); ssh.executeCommandSequence(job.id); const finalJob = await waitForJobCompletion(job.id, config.syncTimeout * params.commands.length); if (finalJob) { return { content: [{ type: "text", text: `Résultat séquence (tâche ${finalJob.id}):\n${JSON.stringify(finalJob, null, 2)}` }] }; } else { return { content: [{ type: "text", text: `Séquence de commandes ${job.id} initiée en arrière-plan.` }] }; } } );
  • Supporting function that implements the sequential SSH command execution logic. Acquires a pooled connection, iterates over commands with per-command timeout/error handling, collects step results, and updates job status accordingly. Called by the tool handler.
    async function executeCommandSequence(jobId) { const job = queue.getJob(jobId); if (!job) return queue.log('error', `Tâche introuvable: ${jobId}`); const results = []; let connection = null; try { const serverConfig = await serverManager.getServer(job.alias); queue.updateJobStatus(jobId, 'running'); // Obtenir une connexion du pool connection = await sshPool.getConnection(job.alias, serverConfig); // Exécuter chaque commande en séquence for (let i = 0; i < job.commands.length; i++) { const cmd = job.commands[i]; const stepJob = { ...job, cmd: typeof cmd === 'string' ? cmd : cmd.command, timeout: cmd.timeout || job.timeout, continueOnError: cmd.continueOnError || job.continueOnError }; try { queue.log('info', `Exécution étape ${i+1}/${job.commands.length}: ${stepJob.cmd}`); const result = await executeWithPooledConnection(connection, stepJob, `${jobId}_step_${i}`); results.push({ step: i + 1, command: stepJob.cmd, success: true, ...result }); } catch (err) { results.push({ step: i + 1, command: stepJob.cmd, success: false, error: err.message }); if (!stepJob.continueOnError) { throw new Error(`Échec à l'étape ${i+1}: ${err.message}`); } } } queue.updateJobStatus(jobId, 'completed', { results }); } catch (err) { queue.updateJobStatus(jobId, 'failed', { error: err.message, results }); } finally { if (connection) { sshPool.releaseConnection(connection.id); } } }

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/fkom13/mcp-sftp-orchestrator'

If you have feedback or need assistance with the MCP directory API, please join our Discord server