sidecar-cli 0.1.2 → 0.1.3-beta.1

package/dist/cli.js

@@ -1,6 +1,7 @@
 #!/usr/bin/env node
 import fs from 'node:fs';
 import path from 'node:path';
+import readline from 'node:readline/promises';
 import { Command } from 'commander';
 import Database from 'better-sqlite3';
 import { z } from 'zod';
@@ -19,15 +20,24 @@ import { buildContext } from './services/context-service.js';
 import { getCapabilitiesManifest } from './services/capabilities-service.js';
 import { addArtifact, listArtifacts } from './services/artifact-service.js';
 import { addDecision, addNote, addWorklog, getActiveSessionId, listRecentEvents } from './services/event-service.js';
-import { addTask, listTasks, markTaskDone } from './services/task-service.js';
 import { currentSession, endSession, startSession, verifySessionHygiene } from './services/session-service.js';
 import { eventIngestSchema, ingestEvent } from './services/event-ingest-service.js';
 import { buildExportJson, buildExportJsonlEvents, writeOutputFile } from './services/export-service.js';
+import { createTaskPacketRecord, getTaskPacket, listTaskPackets } from './tasks/task-service.js';
+import { taskPacketPrioritySchema, taskPacketStatusSchema, taskPacketTypeSchema } from './tasks/task-packet.js';
+import { getRunRecord, listRunRecords, listRunRecordsForTask } from './runs/run-service.js';
+import { runStatusSchema, runnerTypeSchema } from './runs/run-record.js';
+import { compileTaskPrompt } from './prompts/prompt-service.js';
+import { runTaskExecution } from './services/run-orchestrator-service.js';
+import { loadRunnerPreferences } from './runners/config.js';
+import { assignTask, queueReadyTasks } from './services/task-orchestration-service.js';
+import { buildReviewSummary, createFollowupTaskFromRun, reviewRun } from './services/run-review-service.js';
 const pkg = JSON.parse(fs.readFileSync(new URL('../package.json', import.meta.url), 'utf8'));
 const actorSchema = z.enum(['human', 'agent']);
-const taskPrioritySchema = z.enum(['low', 'medium', 'high']);
 const artifactKindSchema = z.enum(['file', 'doc', 'screenshot', 'other']);
-const taskStatusSchema = z.enum(['open', 'done', 'all']);
+const taskListStatusSchema = z.enum(['draft', 'ready', 'queued', 'running', 'review', 'blocked', 'done', 'all']);
+const runListStatusSchema = runStatusSchema.or(z.literal('all'));
+const agentRoleSchema = z.enum(['planner', 'builder-ui', 'builder-app', 'reviewer', 'tester']);
 const exportFormatSchema = z.enum(['json', 'jsonl']);
 const NOT_INITIALIZED_MSG = 'Sidecar is not initialized in this directory or any parent directory';
 function fail(message) {
@@ -185,6 +195,21 @@ function resolveProjectRoot(projectPath) {
     }
     return root;
 }
+function parseCsvOption(input) {
+    if (!input)
+        return [];
+    return input
+        .split(',')
+        .map((item) => item.trim())
+        .filter(Boolean);
+}
+async function askWithDefault(rl, question, fallback) {
+    const suffix = fallback ? ` [${fallback}]` : '';
+    const answer = (await rl.question(`${question}${suffix}: `)).trim();
+    if (answer.length > 0)
+        return answer;
+    return fallback ?? '';
+}
 const program = new Command();
 program.name('sidecar').description('Local-first project memory and recording CLI').version(pkg.version);
 program.option('--no-banner', 'Disable Sidecar banner output');
@@ -275,6 +300,9 @@ program
             fail('Sidecar is already initialized in this project. Re-run with --force to recreate .sidecar files.');
         }
         const files = [
+            sidecar.tasksPath,
+            sidecar.runsPath,
+            sidecar.promptsPath,
             sidecar.dbPath,
             sidecar.configPath,
             sidecar.preferencesPath,
@@ -297,9 +325,12 @@ program
                 sidecar.preferencesPath,
                 sidecar.agentsPath,
                 sidecar.summaryPath,
+                sidecar.tasksPath,
+                sidecar.runsPath,
+                sidecar.promptsPath,
             ]) {
                 if (fs.existsSync(file))
-                    fs.rmSync(file);
+                    fs.rmSync(file, { recursive: true, force: true });
             }
         }
         const db = new Database(sidecar.dbPath);
@@ -315,9 +346,17 @@ program
             settings: {},
         };
         fs.writeFileSync(sidecar.configPath, stringifyJson(config));
+        fs.mkdirSync(sidecar.tasksPath, { recursive: true });
+        fs.mkdirSync(sidecar.runsPath, { recursive: true });
+        fs.mkdirSync(sidecar.promptsPath, { recursive: true });
         fs.writeFileSync(sidecar.preferencesPath, stringifyJson({
             summary: { format: 'markdown', recentLimit: 8 },
             output: { humanTime: true },
+            runner: {
+                defaultRunner: 'codex',
+                preferredRunners: ['codex', 'claude'],
+                defaultAgentRole: 'builder-app',
+            },
         }));
         fs.writeFileSync(sidecar.agentsPath, renderAgentsMarkdown(projectName));
         if (shouldWriteRootAgents) {
@@ -705,46 +744,98 @@ worklog
 });
 const task = program.command('task').description('Task commands');
 task
-    .command('add <title>')
-    .description('Create an open task')
-    .option('--description <text>', 'Description')
+    .command('create')
+    .description('Create a structured task packet')
+    .option('--title <title>', 'Task title')
+    .option('--type <type>', 'feature|bug|chore|research', 'chore')
+    .option('--status <status>', 'draft|ready|queued|running|review|blocked|done', 'draft')
     .option('--priority <priority>', 'low|medium|high', 'medium')
-    .option('--by <actor>', 'human|agent', 'human')
+    .option('--summary <summary>', 'Task summary')
+    .option('--goal <goal>', 'Task goal')
+    .option('--dependencies <task-ids>', 'Comma-separated dependency task IDs')
+    .option('--tags <tags>', 'Comma-separated tags')
+    .option('--target-areas <areas>', 'Comma-separated target areas')
+    .option('--scope-in <items>', 'Comma-separated in-scope items')
+    .option('--scope-out <items>', 'Comma-separated out-of-scope items')
+    .option('--related-decisions <items>', 'Comma-separated related decision IDs/titles')
+    .option('--related-notes <items>', 'Comma-separated related notes')
+    .option('--files-read <paths>', 'Comma-separated files to read')
+    .option('--files-avoid <paths>', 'Comma-separated files to avoid')
+    .option('--constraint-tech <items>', 'Comma-separated technical constraints')
+    .option('--constraint-design <items>', 'Comma-separated design constraints')
+    .option('--validate-cmds <commands>', 'Comma-separated validation commands')
+    .option('--dod <items>', 'Comma-separated definition-of-done checks')
+    .option('--branch <name>', 'Branch name')
+    .option('--worktree <path>', 'Worktree path')
     .option('--json', 'Print machine-readable JSON output')
-    .addHelpText('after', '\nExamples:\n  $ sidecar task add "Ship v0.1"\n  $ sidecar task add "Add tests" --priority high --by agent')
-    .action((title, opts) => {
-    const command = 'task add';
+    .addHelpText('after', '\nExamples:\n  $ sidecar task create\n  $ sidecar task create --title "Add import support" --summary "Support JSON import" --goal "Enable scripted import flow" --priority high\n  $ sidecar task create --title "Refactor parser" --files-read src/parser.ts,src/types.ts --dod "Tests pass,Docs updated"')
+    .action(async (opts) => {
+    const command = 'task create';
     try {
-        const priority = taskPrioritySchema.parse(opts.priority);
-        const by = actorSchema.parse(opts.by);
-        const { db, projectId } = requireInitialized();
-        const result = addTask(db, { projectId, title, description: opts.description, priority, by });
-        db.close();
-        respondSuccess(command, Boolean(opts.json), { task: { id: result.taskId, title, description: opts.description ?? null, status: 'open', priority }, event: { id: result.eventId, type: 'task_created' } }, [`Added task #${result.taskId}.`]);
+        const rootPath = resolveProjectRoot();
+        let title = opts.title?.trim() ?? '';
+        let summary = opts.summary?.trim() ?? '';
+        let goal = opts.goal?.trim() ?? '';
+        if (!title || !summary || !goal) {
+            if (!process.stdin.isTTY) {
+                fail('Missing required fields. Provide --title, --summary, and --goal when not running interactively.');
+            }
+            const rl = readline.createInterface({ input: process.stdin, output: process.stdout });
+            try {
+                title = title || (await askWithDefault(rl, 'Title'));
+                summary = summary || (await askWithDefault(rl, 'Summary', title));
+                goal = goal || (await askWithDefault(rl, 'Goal', `Complete: ${title}`));
+            }
+            finally {
+                rl.close();
+            }
+        }
+        const type = taskPacketTypeSchema.parse(opts.type);
+        const status = taskPacketStatusSchema.parse(opts.status);
+        const priority = taskPacketPrioritySchema.parse(opts.priority);
+        const created = createTaskPacketRecord(rootPath, {
+            title,
+            summary,
+            goal,
+            type,
+            status,
+            priority,
+            scope_in_scope: parseCsvOption(opts.scopeIn),
+            scope_out_of_scope: parseCsvOption(opts.scopeOut),
+            related_decisions: parseCsvOption(opts.relatedDecisions),
+            related_notes: parseCsvOption(opts.relatedNotes),
+            files_to_read: parseCsvOption(opts.filesRead),
+            files_to_avoid: parseCsvOption(opts.filesAvoid),
+            technical_constraints: parseCsvOption(opts.constraintTech),
+            design_constraints: parseCsvOption(opts.constraintDesign),
+            validation_commands: parseCsvOption(opts.validateCmds),
+            dependencies: parseCsvOption(opts.dependencies).map((v) => v.toUpperCase()),
+            tags: parseCsvOption(opts.tags),
+            target_areas: parseCsvOption(opts.targetAreas),
+            definition_of_done: parseCsvOption(opts.dod),
+            branch: opts.branch?.trim(),
+            worktree: opts.worktree?.trim(),
+        });
+        respondSuccess(command, Boolean(opts.json), { task: created.task, path: created.path }, [`Created task ${created.task.task_id}.`, `Path: ${created.path}`]);
     }
     catch (err) {
         handleCommandError(command, Boolean(opts.json), err);
     }
 });
 task
-    .command('done <task-id>')
-    .description('Mark a task as done')
-    .option('--by <actor>', 'human|agent', 'human')
+    .command('show <task-id>')
+    .description('Show a task packet by id')
     .option('--json', 'Print machine-readable JSON output')
-    .addHelpText('after', '\nExamples:\n  $ sidecar task done 3\n  $ sidecar task done 3 --json')
+    .addHelpText('after', '\nExamples:\n  $ sidecar task show T-001\n  $ sidecar task show T-001 --json')
     .action((taskIdText, opts) => {
-    const command = 'task done';
+    const command = 'task show';
     try {
-        const taskId = Number.parseInt(taskIdText, 10);
-        if (!Number.isInteger(taskId) || taskId <= 0)
-            fail('Task id must be a positive integer');
-        const by = actorSchema.parse(opts.by);
-        const { db, projectId } = requireInitialized();
-        const result = markTaskDone(db, { projectId, taskId, by });
-        db.close();
-        if (!result.ok)
-            fail(result.reason);
-        respondSuccess(command, Boolean(opts.json), { task: { id: taskId, status: 'done' }, event: { id: result.eventId, type: 'task_completed' } }, [`Completed task #${taskId}.`]);
+        const task = getTaskPacket(resolveProjectRoot(), taskIdText.trim().toUpperCase());
+        if (opts.json) {
+            respondSuccess(command, true, { task }, []);
+            return;
+        }
+        console.log(stringifyJson(task));
     }
     catch (err) {
         handleCommandError(command, Boolean(opts.json), err);
@@ -752,40 +843,327 @@ task
 });
 task
     .command('list')
-    .description('List tasks by status')
-    .option('--status <status>', 'open|done|all', 'open')
-    .option('--format <format>', 'table|json', 'table')
+    .description('List task packets')
+    .option('--status <status>', 'draft|ready|queued|running|review|blocked|done|all', 'all')
     .option('--json', 'Print machine-readable JSON output')
-    .addHelpText('after', '\nExamples:\n  $ sidecar task list\n  $ sidecar task list --status all --format json')
+    .addHelpText('after', '\nExamples:\n  $ sidecar task list\n  $ sidecar task list --status open\n  $ sidecar task list --json')
     .action((opts) => {
     const command = 'task list';
     try {
-        const status = taskStatusSchema.parse(opts.status);
-        const { db, projectId } = requireInitialized();
-        const rows = listTasks(db, { projectId, status });
-        db.close();
-        if (opts.format === 'json' || opts.json) {
-            if (opts.json)
-                printJsonEnvelope(jsonSuccess(command, { status, tasks: rows }));
-            else
-                console.log(stringifyJson(rows));
+        const status = taskListStatusSchema.parse(opts.status);
+        const tasks = listTaskPackets(resolveProjectRoot());
+        const rows = status === 'all' ? tasks : tasks.filter((task) => task.status === status);
+        if (opts.json) {
+            respondSuccess(command, true, { status, tasks: rows }, []);
             return;
         }
-        const taskRows = rows;
-        if (taskRows.length === 0) {
+        if (rows.length === 0) {
             console.log('No tasks found.');
             return;
         }
-        const idWidth = Math.max(2, ...taskRows.map((r) => String(r.id).length));
-        const statusWidth = Math.max(6, ...taskRows.map((r) => r.status.length));
-        const priorityWidth = Math.max(8, ...taskRows.map((r) => (r.priority ?? 'n/a').length));
-        console.log(`${'ID'.padEnd(idWidth)}  ${'STATUS'.padEnd(statusWidth)}  ${'PRIORITY'.padEnd(priorityWidth)}  TITLE`);
-        for (const row of taskRows) {
-            const id = String(row.id).padEnd(idWidth);
-            const statusLabel = row.status.padEnd(statusWidth);
-            const prio = (row.priority ?? 'n/a').padEnd(priorityWidth);
-            console.log(`${id}  ${statusLabel}  ${prio}  ${row.title}`);
+        const idWidth = Math.max(6, ...rows.map((r) => r.task_id.length));
+        const statusWidth = Math.max(11, ...rows.map((r) => r.status.length));
+        const priorityWidth = Math.max(8, ...rows.map((r) => r.priority.length));
+        console.log(`${'TASK ID'.padEnd(idWidth)}  ${'STATUS'.padEnd(statusWidth)}  ${'PRIORITY'.padEnd(priorityWidth)}  TITLE`);
+        for (const row of rows) {
+            console.log(`${row.task_id.padEnd(idWidth)}  ${row.status.padEnd(statusWidth)}  ${row.priority.padEnd(priorityWidth)}  ${row.title}`);
+        }
+    }
+    catch (err) {
+        handleCommandError(command, Boolean(opts.json), err);
+    }
+});
+task
+    .command('assign <task-id>')
+    .description('Auto-assign agent role and runner for a task')
+    .option('--agent-role <role>', 'planner|builder-ui|builder-app|reviewer|tester')
+    .option('--runner <runner>', 'codex|claude')
+    .option('--json', 'Print machine-readable JSON output')
+    .addHelpText('after', '\nExamples:\n  $ sidecar task assign T-001\n  $ sidecar task assign T-001 --agent-role builder-ui --runner codex\n  $ sidecar task assign T-001 --json')
+    .action((taskIdText, opts) => {
+    const command = 'task assign';
+    try {
+        const rootPath = resolveProjectRoot();
+        const result = assignTask(rootPath, taskIdText.trim().toUpperCase(), {
+            role: opts.agentRole ? agentRoleSchema.parse(opts.agentRole) : undefined,
+            runner: opts.runner ? runnerTypeSchema.parse(opts.runner) : undefined,
+        });
+        respondSuccess(command, Boolean(opts.json), result, [
+            `Assigned ${result.task_id}.`,
+            `Role: ${result.agent_role}`,
+            `Runner: ${result.runner}`,
+            `Reason: ${result.reason}`,
+        ]);
+    }
+    catch (err) {
+        handleCommandError(command, Boolean(opts.json), err);
+    }
+});
+task
+    .command('create-followup <run-id>')
+    .description('Create a follow-up task packet from a run report')
+    .option('--json', 'Print machine-readable JSON output')
+    .addHelpText('after', '\nExamples:\n  $ sidecar task create-followup R-010\n  $ sidecar task create-followup R-010 --json')
+    .action((runIdText, opts) => {
+    const command = 'task create-followup';
+    try {
+        const result = createFollowupTaskFromRun(resolveProjectRoot(), runIdText.trim().toUpperCase());
+        respondSuccess(command, Boolean(opts.json), result, [
+            `Created follow-up task ${result.task_id}.`,
+            `Source run: ${result.source_run_id}`,
+            `Title: ${result.title}`,
+        ]);
+    }
+    catch (err) {
+        handleCommandError(command, Boolean(opts.json), err);
+    }
+});
+const prompt = program.command('prompt').description('Prompt compilation commands');
+prompt
+    .command('compile <task-id>')
+    .description('Compile a markdown execution brief from a task packet')
+    .requiredOption('--runner <runner>', 'codex|claude')
+    .requiredOption('--agent-role <role>', 'Agent role, for example builder')
+    .option('--preview', 'Print compiled prompt content after writing file')
+    .option('--json', 'Print machine-readable JSON output')
+    .addHelpText('after', '\nExamples:\n  $ sidecar prompt compile T-001 --runner codex --agent-role builder\n  $ sidecar prompt compile T-001 --runner claude --agent-role builder --preview\n  $ sidecar prompt compile T-001 --runner codex --agent-role reviewer --json')
+    .action((taskIdText, opts) => {
+    const command = 'prompt compile';
+    try {
+        const rootPath = resolveProjectRoot();
+        const taskId = taskIdText.trim().toUpperCase();
+        const runner = runnerTypeSchema.parse(opts.runner);
+        const agentRole = String(opts.agentRole ?? '').trim();
+        if (!agentRole)
+            fail('Agent role is required');
+        const compiled = compileTaskPrompt({
+            rootPath,
+            taskId,
+            runner,
+            agentRole,
+        });
+        respondSuccess(command, Boolean(opts.json), {
+            run_id: compiled.run_id,
+            task_id: compiled.task_id,
+            runner_type: compiled.runner_type,
+            agent_role: compiled.agent_role,
+            prompt_path: compiled.prompt_path,
+            preview: opts.preview ? compiled.prompt_markdown : null,
+        }, [
+            `Compiled prompt for ${compiled.task_id}.`,
+            `Run: ${compiled.run_id}`,
+            `Path: ${compiled.prompt_path}`,
+            ...(opts.preview ? ['', compiled.prompt_markdown] : []),
+        ]);
+    }
+    catch (err) {
+        handleCommandError(command, Boolean(opts.json), err);
+    }
+});
+program
+    .command('run-exec <task-id>')
+    .description('Internal command backing `sidecar run <task-id>`')
+    .option('--runner <runner>', 'codex|claude')
+    .option('--agent-role <role>', 'planner|builder-ui|builder-app|reviewer|tester')
+    .option('--dry-run', 'Prepare and compile only without executing external runner')
+    .option('--json', 'Print machine-readable JSON output')
+    .action((taskIdText, opts) => {
+    const command = 'run';
+    try {
+        const rootPath = resolveProjectRoot();
+        const defaults = loadRunnerPreferences(rootPath);
+        const selectedRunner = opts.runner ? runnerTypeSchema.parse(opts.runner) : defaults.default_runner;
+        const selectedAgentRole = opts.agentRole
+            ? agentRoleSchema.parse(opts.agentRole)
+            : defaults.default_agent_role;
+        const result = runTaskExecution({
+            rootPath,
+            taskId: String(taskIdText).trim().toUpperCase(),
+            runner: selectedRunner,
+            agentRole: selectedAgentRole,
+            dryRun: Boolean(opts.dryRun),
+        });
+        respondSuccess(command, Boolean(opts.json), result, [
+            `Prepared run ${result.run_id} for ${result.task_id}.`,
+            `Runner: ${result.runner_type} (${result.agent_role})`,
+            `Prompt: ${result.prompt_path}`,
+            `Command: ${result.shell_command}`,
+            `Status: ${result.status}`,
+            `Summary: ${result.summary}`,
+        ]);
+    }
+    catch (err) {
+        handleCommandError(command, Boolean(opts.json), err);
+    }
+});
+const run = program
+    .command('run')
+    .description('Run task execution or inspect run records')
+    .addHelpText('after', '\nExamples:\n  $ sidecar run T-001 --dry-run\n  $ sidecar run T-001 --runner claude --agent-role reviewer\n  $ sidecar run queue\n  $ sidecar run start-ready --dry-run\n  $ sidecar run list --task T-001\n  $ sidecar run show R-001');
+run
+    .command('queue')
+    .description('Queue all ready tasks with satisfied dependencies')
+    .option('--json', 'Print machine-readable JSON output')
+    .addHelpText('after', '\nExamples:\n  $ sidecar run queue\n  $ sidecar run queue --json')
+    .action((opts) => {
+    const command = 'run queue';
+    try {
+        const rootPath = resolveProjectRoot();
+        const decisions = queueReadyTasks(rootPath);
+        respondSuccess(command, Boolean(opts.json), { decisions }, [
+            `Processed ${decisions.length} ready task(s).`,
+            ...decisions.map((d) => `- ${d.task_id}: ${d.reason}`),
+        ]);
+    }
+    catch (err) {
+        handleCommandError(command, Boolean(opts.json), err);
+    }
+});
+run
+    .command('start-ready')
+    .description('Queue and start all runnable ready tasks')
+    .option('--dry-run', 'Prepare and compile only without executing external runners')
+    .option('--json', 'Print machine-readable JSON output')
+    .addHelpText('after', '\nExamples:\n  $ sidecar run start-ready\n  $ sidecar run start-ready --dry-run --json')
+    .action((opts) => {
+    const command = 'run start-ready';
+    try {
+        const rootPath = resolveProjectRoot();
+        const queueDecisions = queueReadyTasks(rootPath);
+        const queuedTasks = listTaskPackets(rootPath).filter((task) => task.status === 'queued');
+        const results = queuedTasks.map((task) => runTaskExecution({ rootPath, taskId: task.task_id, dryRun: Boolean(opts.dryRun) }));
+        respondSuccess(command, Boolean(opts.json), { queued: queueDecisions, results }, [
+            `Queued in this pass: ${queueDecisions.filter((d) => d.queued).length}`,
+            `Started: ${results.length}`,
+            ...results.map((r) => `- ${r.task_id} -> ${r.run_id} (${r.status})`),
+        ]);
+    }
+    catch (err) {
+        handleCommandError(command, Boolean(opts.json), err);
+    }
+});
+run
+    .command('approve <run-id>')
+    .description('Review a completed run as approved, needs changes, or merged')
+    .option('--state <state>', 'approved|needs_changes|merged', 'approved')
+    .option('--note <text>', 'Review note')
+    .option('--by <name>', 'Reviewer name', 'human')
+    .option('--json', 'Print machine-readable JSON output')
+    .addHelpText('after', '\nExamples:\n  $ sidecar run approve R-010\n  $ sidecar run approve R-010 --state needs_changes --note "Address test failures"\n  $ sidecar run approve R-010 --state merged --json')
+    .action((runIdText, opts) => {
+    const command = 'run approve';
+    try {
+        const state = String(opts.state);
+        if (state !== 'approved' && state !== 'needs_changes' && state !== 'merged') {
+            fail('State must be one of: approved, needs_changes, merged');
+        }
+        const result = reviewRun(resolveProjectRoot(), runIdText.trim().toUpperCase(), state, {
+            note: opts.note,
+            by: opts.by,
+        });
+        respondSuccess(command, Boolean(opts.json), result, [
+            `Run ${result.run_id} marked ${result.review_state}.`,
+            `Task ${result.task_id} -> ${result.task_status}`,
+        ]);
+    }
+    catch (err) {
+        handleCommandError(command, Boolean(opts.json), err);
+    }
+});
+run
+    .command('block <run-id>')
+    .description('Mark a completed run as blocked and set linked task blocked')
+    .option('--note <text>', 'Blocking reason')
+    .option('--by <name>', 'Reviewer name', 'human')
+    .option('--json', 'Print machine-readable JSON output')
+    .addHelpText('after', '\nExamples:\n  $ sidecar run block R-010 --note "Migration failed"\n  $ sidecar run block R-010 --json')
+    .action((runIdText, opts) => {
+    const command = 'run block';
+    try {
+        const result = reviewRun(resolveProjectRoot(), runIdText.trim().toUpperCase(), 'blocked', {
+            note: opts.note,
+            by: opts.by,
+        });
+        respondSuccess(command, Boolean(opts.json), result, [
+            `Run ${result.run_id} marked blocked.`,
+            `Task ${result.task_id} -> ${result.task_status}`,
+        ]);
+    }
+    catch (err) {
+        handleCommandError(command, Boolean(opts.json), err);
+    }
+});
+run
+    .command('list')
+    .description('List execution run records')
+    .option('--task <task-id>', 'Filter by task id (for example T-001)')
+    .option('--status <status>', 'queued|preparing|running|review|blocked|completed|failed|all', 'all')
+    .option('--json', 'Print machine-readable JSON output')
+    .addHelpText('after', '\nExamples:\n  $ sidecar run list\n  $ sidecar run list --task T-001\n  $ sidecar run list --status completed --json')
+    .action((opts) => {
+    const command = 'run list';
+    try {
+        const rootPath = resolveProjectRoot();
+        const status = runListStatusSchema.parse(opts.status);
+        const base = opts.task ? listRunRecordsForTask(rootPath, String(opts.task).trim().toUpperCase()) : listRunRecords(rootPath);
+        const rows = status === 'all' ? base : base.filter((entry) => entry.status === status);
+        if (opts.json) {
+            respondSuccess(command, true, { status, task_id: opts.task ? String(opts.task).trim().toUpperCase() : null, runs: rows }, []);
+            return;
+        }
+        if (rows.length === 0) {
+            console.log('No run records found.');
+            return;
+        }
+        const idWidth = Math.max(6, ...rows.map((r) => r.run_id.length));
+        const taskWidth = Math.max(7, ...rows.map((r) => r.task_id.length));
+        const statusWidth = Math.max(10, ...rows.map((r) => r.status.length));
+        console.log(`${'RUN ID'.padEnd(idWidth)}  ${'TASK ID'.padEnd(taskWidth)}  ${'STATUS'.padEnd(statusWidth)}  STARTED`);
+        for (const row of rows) {
+            console.log(`${row.run_id.padEnd(idWidth)}  ${row.task_id.padEnd(taskWidth)}  ${row.status.padEnd(statusWidth)}  ${humanTime(row.started_at)}`);
+        }
+    }
+    catch (err) {
+        handleCommandError(command, Boolean(opts.json), err);
+    }
+});
+run
+    .command('summary')
+    .description('Show project-level run review summary')
+    .option('--json', 'Print machine-readable JSON output')
+    .addHelpText('after', '\nExamples:\n  $ sidecar run summary\n  $ sidecar run summary --json')
+    .action((opts) => {
+    const command = 'run summary';
+    try {
+        const data = buildReviewSummary(resolveProjectRoot());
+        respondSuccess(command, Boolean(opts.json), data, [
+            `Completed runs: ${data.completed_runs}`,
+            `Blocked runs: ${data.blocked_runs}`,
+            `Suggested follow-ups: ${data.suggested_follow_ups}`,
+            'Recently merged:',
+            ...(data.recently_merged.length
+                ? data.recently_merged.map((r) => `- ${r.run_id} (${r.task_id}) at ${humanTime(r.reviewed_at)}`)
+                : ['- none']),
+        ]);
+    }
+    catch (err) {
+        handleCommandError(command, Boolean(opts.json), err);
+    }
+});
+run
+    .command('show <run-id>')
+    .description('Show a run record by id')
+    .option('--json', 'Print machine-readable JSON output')
+    .addHelpText('after', '\nExamples:\n  $ sidecar run show R-001\n  $ sidecar run show R-001 --json')
+    .action((runIdText, opts) => {
+    const command = 'run show';
+    try {
+        const runRecord = getRunRecord(resolveProjectRoot(), runIdText.trim().toUpperCase());
+        if (opts.json) {
+            respondSuccess(command, true, { run: runRecord }, []);
+            return;
         }
+        console.log(stringifyJson(runRecord));
     }
     catch (err) {
         handleCommandError(command, Boolean(opts.json), err);
@@ -952,5 +1330,17 @@ if (process.argv.length === 2) {
     maybePrintUpdateNotice();
     process.exit(0);
 }
+if (process.argv[2] === 'run' &&
+    process.argv[3] &&
+    !process.argv[3].startsWith('-') &&
+    process.argv[3] !== 'list' &&
+    process.argv[3] !== 'show' &&
+    process.argv[3] !== 'queue' &&
+    process.argv[3] !== 'start-ready' &&
+    process.argv[3] !== 'approve' &&
+    process.argv[3] !== 'block' &&
+    process.argv[3] !== 'summary') {
+    process.argv.splice(2, 1, 'run-exec');
+}
 program.parse(process.argv);
 maybePrintUpdateNotice();

package/dist/lib/paths.js

@@ -6,6 +6,9 @@ export function getSidecarPaths(rootPath) {
     return {
         rootPath,
         sidecarPath,
+        tasksPath: path.join(sidecarPath, 'tasks'),
+        runsPath: path.join(sidecarPath, 'runs'),
+        promptsPath: path.join(sidecarPath, 'prompts'),
         rootAgentsPath: path.join(rootPath, 'AGENTS.md'),
         rootClaudePath: path.join(rootPath, 'CLAUDE.md'),
         dbPath: path.join(sidecarPath, 'sidecar.db'),

package/dist/lib/ui.js

@@ -84,7 +84,10 @@ export function launchUiServer(options) {
     }
     const child = spawn(process.execPath, [serverPath, '--project', options.projectPath, '--port', String(options.port)], {
         stdio: 'inherit',
-        env: { ...process.env },
+        env: {
+            ...process.env,
+            SIDECAR_CLI_JS: process.argv[1] || '',
+        },
     });
     const url = `http://localhost:${options.port}`;
     if (options.openBrowser) {