feat(cli): Add replay command to view results from JSONL log files

src/cli/args.test.ts

@@ -344,6 +344,55 @@ describe('parseCliArgs', () => {
     expect(result.command).toBe('sync');
     expect(result.options.remote).toBe('getsentry/skills');
   });
+
+  it('parses replay command', () => {
+    const result = parseCliArgs(['replay']);
+    expect(result.command).toBe('replay');
+    expect(result.replayOptions).toBeDefined();
+    expect(result.replayOptions!.files).toEqual([]);
+  });
+
+  it('parses replay command with single file', () => {
+    const result = parseCliArgs(['replay', 'run.jsonl']);
+    expect(result.command).toBe('replay');
+    expect(result.replayOptions!.files).toEqual(['run.jsonl']);
+  });
+
+  it('parses replay command with multiple files', () => {
+    const result = parseCliArgs(['replay', 'run1.jsonl', 'run2.jsonl', 'run3.jsonl']);
+    expect(result.command).toBe('replay');
+    expect(result.replayOptions!.files).toEqual(['run1.jsonl', 'run2.jsonl', 'run3.jsonl']);
+  });
+
+  it('parses replay command with --json flag', () => {
+    const result = parseCliArgs(['replay', 'run.jsonl', '--json']);
+    expect(result.command).toBe('replay');
+    expect(result.options.json).toBe(true);
+  });
+
+  it('parses replay command with --report-on option', () => {
+    const result = parseCliArgs(['replay', 'run.jsonl', '--report-on', 'high']);
+    expect(result.command).toBe('replay');
+    expect(result.options.reportOn).toBe('high');
+  });
+
+  it('parses replay command with --min-confidence option', () => {
+    const result = parseCliArgs(['replay', 'run.jsonl', '--min-confidence', 'high']);
+    expect(result.command).toBe('replay');
+    expect(result.options.minConfidence).toBe('high');
+  });
+
+  it('parses replay command with verbosity flags', () => {
+    const result = parseCliArgs(['replay', 'run.jsonl', '-v']);
+    expect(result.command).toBe('replay');
+    expect(result.options.verbose).toBe(1);
+  });
+
+  it('parses replay command with --quiet flag', () => {
+    const result = parseCliArgs(['replay', 'run.jsonl', '--quiet']);
+    expect(result.command).toBe('replay');
+    expect(result.options.quiet).toBe(true);
+  });
 });
 
 describe('CLIOptionsSchema', () => {

src/cli/args.ts

@@ -55,10 +55,15 @@ export interface SetupAppOptions {
   open: boolean;
 }
 
+export interface ReplayOptions {
+  files: string[];
+}
+
 export interface ParsedArgs {
-  command: 'run' | 'help' | 'init' | 'add' | 'version' | 'setup-app' | 'sync';
+  command: 'run' | 'help' | 'init' | 'add' | 'version' | 'setup-app' | 'sync' | 'replay';
   options: CLIOptions;
   setupAppOptions?: SetupAppOptions;
+  replayOptions?: ReplayOptions;
 }
 
 export function showVersion(): void {
@@ -75,6 +80,7 @@ Commands:
   add [skill]          Add a skill trigger to warden.toml
   sync [remote]        Update cached remote skills to latest
   setup-app            Create a GitHub App for Warden via manifest flow
+  replay <files...>    Replay results from JSONL log files
   (default)            Run analysis on targets or using warden.toml triggers
 
 Targets:
@@ -313,7 +319,7 @@ export function parseCliArgs(argv: string[] = process.argv.slice(2)): ParsedArgs
   }
 
   // Filter out known commands from positionals
-  const commands = ['run', 'help', 'init', 'add', 'version', 'setup-app', 'sync'];
+  const commands = ['run', 'help', 'init', 'add', 'version', 'setup-app', 'sync', 'replay'];
   const targets = positionals.filter((p) => !commands.includes(p));
 
   // Handle explicit help command
@@ -397,6 +403,30 @@ export function parseCliArgs(argv: string[] = process.argv.slice(2)): ParsedArgs
     };
   }
 
+  // Handle replay command
+  if (positionals.includes('replay')) {
+    // All positionals after 'replay' are log files
+    const replayIndex = positionals.indexOf('replay');
+    const logFiles = positionals.slice(replayIndex + 1);
+
+    return {
+      command: 'replay',
+      options: CLIOptionsSchema.parse({
+        json: values.json,
+        reportOn: values['report-on'] as SeverityThreshold | undefined,
+        minConfidence: values['min-confidence'] as ConfidenceThreshold | undefined,
+        quiet: values.quiet,
+        verbose: verboseCount,
+        debug: values.debug,
+        log: values.log,
+        color: resolveColorOption(values),
+      }),
+      replayOptions: {
+        files: logFiles,
+      },
+    };
+  }
+
   const rawOptions = {
     targets: targets.length > 0 ? targets : undefined,
     skill: values.skill,

src/cli/main.ts

@@ -1,4 +1,4 @@
-import { existsSync } from 'node:fs';
+import { existsSync, readFileSync } from 'node:fs';
 import { dirname, join, resolve } from 'node:path';
 import { config as dotenvConfig } from 'dotenv';
 import { Sentry, flushSentry, setGlobalAttributes, emitRunMetric, getTraceId } from '../sentry.js';
@@ -9,7 +9,7 @@ import { matchTrigger, filterContextByPaths, shouldFail, countFindingsAtOrAbove
 import type { SkillReport, ConfidenceThreshold } from '../types/index.js';
 import { filterFindings } from '../types/index.js';
 import { DEFAULT_CONCURRENCY, getAnthropicApiKey } from '../utils/index.js';
-import { parseCliArgs, showHelp, showVersion, classifyTargets, type CLIOptions } from './args.js';
+import { parseCliArgs, showHelp, showVersion, classifyTargets, type CLIOptions, type ReplayOptions } from './args.js';
 import { buildLocalEventContext, buildFileEventContext } from './context.js';
 import { getRepoRoot, refExists, hasUncommittedChanges } from './git.js';
 import { renderTerminalReport, filterReports } from './terminal.js';
@@ -25,6 +25,8 @@ import {
   renderJsonlString,
   getRepoLogPath,
   generateRunId,
+  parseJsonlReports,
+  type JsonlRunMetadata,
   type SkillTaskOptions,
 } from './output/index.js';
 import { cleanupLogs } from './log-cleanup.js';
@@ -718,6 +720,84 @@ async function runDirectSkillMode(options: CLIOptions, reporter: Reporter): Prom
   return runSkills(context, options, reporter);
 }
 
+/**
+ * Run in replay mode: render results from JSONL log files.
+ */
+async function runReplay(replayOptions: ReplayOptions, options: CLIOptions, reporter: Reporter): Promise<number> {
+  const { files } = replayOptions;
+
+  if (files.length === 0) {
+    reporter.error('No log files specified');
+    reporter.tip('Usage: warden replay <file.jsonl> [file2.jsonl ...]');
+    return 1;
+  }
+
+  // Validate all files exist before processing
+  const missingFiles: string[] = [];
+  for (const file of files) {
+    if (!existsSync(file)) {
+      missingFiles.push(file);
+    }
+  }
+
+  if (missingFiles.length > 0) {
+    reporter.error(`Log ${pluralize(missingFiles.length, 'file')} not found: ${missingFiles.join(', ')}`);
+    return 1;
+  }
+
+  // Parse and merge reports from all files
+  const allReports: SkillReport[] = [];
+  let totalDurationMs = 0;
+  let lastRunMetadata: JsonlRunMetadata | undefined;
+
+  for (const file of files) {
+    try {
+      const content = readFileSync(file, 'utf-8');
+      const parsed = parseJsonlReports(content);
+      allReports.push(...parsed.reports);
+      totalDurationMs = Math.max(totalDurationMs, parsed.totalDurationMs);
+
+      if (parsed.runMetadata) {
+        lastRunMetadata = parsed.runMetadata;
+        reporter.debug(`Loaded ${parsed.reports.length} ${pluralize(parsed.reports.length, 'skill')} from ${file}`);
+        reporter.debug(`  Run ID: ${parsed.runMetadata.runId}`);
+        reporter.debug(`  Timestamp: ${parsed.runMetadata.timestamp}`);
+      }
+    } catch (err) {
+      reporter.error(`Failed to parse ${file}: ${err instanceof Error ? err.message : String(err)}`);
+      return 1;
+    }
+  }
+
+  if (allReports.length === 0) {
+    reporter.warning('No skill reports found in log files');
+    return 0;
+  }
+
+  // Apply filtering
+  const filteredReports = filterReports(allReports, options.reportOn, options.minConfidence ?? 'medium');
+
+  // Output results
+  reporter.blank();
+  if (options.json) {
+    // Re-render as JSONL for piping, preserving original run metadata
+    const jsonlContent = renderJsonlString(filteredReports, totalDurationMs, lastRunMetadata ? {
+      runId: lastRunMetadata.runId,
+      traceId: lastRunMetadata.traceId,
+      timestamp: new Date(lastRunMetadata.timestamp),
+    } : undefined);
+    process.stdout.write(jsonlContent);
+  } else {
+    console.log(renderTerminalReport(filteredReports, reporter.mode, { verbosity: reporter.verbosity }));
+  }
+
+  // Show summary
+  reporter.blank();
+  reporter.renderSummary(filteredReports, totalDurationMs);
+
+  return 0;
+}
+
 async function runCommand(options: CLIOptions, reporter: Reporter): Promise<number> {
   const targets = options.targets ?? [];
 
@@ -759,7 +839,7 @@ async function runCommand(options: CLIOptions, reporter: Reporter): Promise<numb
 }
 
 export async function main(): Promise<void> {
-  const { command, options, setupAppOptions } = parseCliArgs();
+  const { command, options, setupAppOptions, replayOptions } = parseCliArgs();
 
   if (command === 'help') {
     showHelp();
@@ -813,6 +893,12 @@ export async function main(): Promise<void> {
           return runSetupApp(setupAppOptions, reporter);
         case 'sync':
           return runSync(options, reporter);
+        case 'replay':
+          if (!replayOptions) {
+            reporter.error('Missing replay options');
+            process.exit(1);
+          }
+          return runReplay(replayOptions, options, reporter);
         default:
           return runCommand(options, reporter);
       }

src/cli/output/index.ts

@@ -43,7 +43,9 @@ export {
   renderJsonlString,
   getRepoLogPath,
   generateRunId,
+  parseJsonlReports,
   type JsonlRecord,
   type JsonlRunMetadata,
+  type ParsedJsonlLog,
 } from './jsonl.js';
 export { ICON_CHECK, ICON_SKIPPED, SPINNER_FRAMES } from './icons.js';

src/cli/output/jsonl.test.ts

@@ -8,6 +8,8 @@ import {
   generateRunId,
   shortRunId,
   readJsonlLog,
+  parseJsonlReports,
+  renderJsonlString,
   type JsonlRecord,
 } from './jsonl.js';
 import type { SkillReport } from '../../types/index.js';
@@ -551,3 +553,122 @@ describe('repo-local logging integration', () => {
     expect(content2.run.runId).toBe(runId2);
   });
 });
+
+describe('parseJsonlReports', () => {
+  it('reconstructs SkillReport from JSONL content', () => {
+    // Sample JSONL content that matches what would be written by renderJsonlString
+    const jsonlContent = `{"run":{"timestamp":"2026-02-18T14:32:15.123Z","durationMs":2000,"cwd":"/test","runId":"test-123"},"skill":"security-review","summary":"Found 2 issues","findings":[{"id":"sec-001","severity":"high","title":"SQL Injection","description":"User input passed directly to query"},{"id":"sec-002","severity":"medium","title":"XSS Risk","description":"Unescaped output"}],"durationMs":1234,"usage":{"inputTokens":1000,"outputTokens":500,"costUSD":0.01}}
+{"run":{"timestamp":"2026-02-18T14:32:15.123Z","durationMs":2000,"cwd":"/test","runId":"test-123"},"type":"summary","totalFindings":2,"bySeverity":{"critical":0,"high":1,"medium":1,"low":0,"info":0},"usage":{"inputTokens":1000,"outputTokens":500,"costUSD":0.01}}
+`;
+
+    const result = parseJsonlReports(jsonlContent);
+
+    expect(result.reports.length).toBe(1);
+    expect(result.reports[0]!.skill).toBe('security-review');
+    expect(result.reports[0]!.findings.length).toBe(2);
+    expect(result.reports[0]!.findings[0]!.id).toBe('sec-001');
+    expect(result.reports[0]!.durationMs).toBe(1234);
+    expect(result.reports[0]!.usage?.inputTokens).toBe(1000);
+    expect(result.totalDurationMs).toBe(2000);
+    expect(result.runMetadata?.runId).toBe('test-123');
+  });
+
+  it('handles multiple skill records', () => {
+    const jsonlContent = `{"run":{"timestamp":"2026-02-18T14:32:15.123Z","durationMs":3000,"cwd":"/test","runId":"multi-123"},"skill":"skill-1","summary":"Done","findings":[],"durationMs":1000}
+{"run":{"timestamp":"2026-02-18T14:32:15.123Z","durationMs":3000,"cwd":"/test","runId":"multi-123"},"skill":"skill-2","summary":"Issues found","findings":[{"id":"a","severity":"low","title":"A","description":"A"}],"durationMs":2000}
+{"run":{"timestamp":"2026-02-18T14:32:15.123Z","durationMs":3000,"cwd":"/test","runId":"multi-123"},"type":"summary","totalFindings":1,"bySeverity":{"critical":0,"high":0,"medium":0,"low":1,"info":0}}
+`;
+
+    const result = parseJsonlReports(jsonlContent);
+
+    expect(result.reports.length).toBe(2);
+    expect(result.reports[0]!.skill).toBe('skill-1');
+    expect(result.reports[1]!.skill).toBe('skill-2');
+    expect(result.reports[1]!.findings.length).toBe(1);
+    expect(result.totalDurationMs).toBe(3000);
+  });
+
+  it('handles empty logs (summary only)', () => {
+    const jsonlContent = `{"run":{"timestamp":"2026-02-18T14:32:15.123Z","durationMs":100,"cwd":"/test","runId":"empty-123"},"type":"summary","totalFindings":0,"bySeverity":{"critical":0,"high":0,"medium":0,"low":0,"info":0}}
+`;
+
+    const result = parseJsonlReports(jsonlContent);
+
+    expect(result.reports.length).toBe(0);
+    expect(result.totalDurationMs).toBe(100);
+    expect(result.runMetadata?.runId).toBe('empty-123');
+  });
+
+  it('skips invalid lines gracefully', () => {
+    const jsonlContent = `invalid json here
+{"run":{"timestamp":"2026-02-18T14:32:15.123Z","durationMs":1000,"cwd":"/test","runId":"partial-123"},"skill":"valid-skill","summary":"OK","findings":[]}
+another bad line
+{"run":{"timestamp":"2026-02-18T14:32:15.123Z","durationMs":1000,"cwd":"/test","runId":"partial-123"},"type":"summary","totalFindings":0,"bySeverity":{"critical":0,"high":0,"medium":0,"low":0,"info":0}}
+`;
+
+    const result = parseJsonlReports(jsonlContent);
+
+    expect(result.reports.length).toBe(1);
+    expect(result.reports[0]!.skill).toBe('valid-skill');
+  });
+
+  it('reconstructs files array from JSONL', () => {
+    const jsonlContent = `{"run":{"timestamp":"2026-02-18T14:32:15.123Z","durationMs":2000,"cwd":"/test","runId":"files-123"},"skill":"review","summary":"Done","findings":[],"files":[{"filename":"src/api.ts","findings":1,"durationMs":1200},{"filename":"src/utils.ts","findings":0,"durationMs":800}]}
+{"run":{"timestamp":"2026-02-18T14:32:15.123Z","durationMs":2000,"cwd":"/test","runId":"files-123"},"type":"summary","totalFindings":1,"bySeverity":{"critical":0,"high":1,"medium":0,"low":0,"info":0}}
+`;
+
+    const result = parseJsonlReports(jsonlContent);
+
+    expect(result.reports[0]!.files).toBeDefined();
+    expect(result.reports[0]!.files!.length).toBe(2);
+    expect(result.reports[0]!.files![0]!.filename).toBe('src/api.ts');
+    expect(result.reports[0]!.files![0]!.findingCount).toBe(1);
+    expect(result.reports[0]!.files![1]!.filename).toBe('src/utils.ts');
+  });
+
+  it('handles skippedFiles in reports', () => {
+    const jsonlContent = `{"run":{"timestamp":"2026-02-18T14:32:15.123Z","durationMs":1000,"cwd":"/test","runId":"skip-123"},"skill":"review","summary":"Done","findings":[],"skippedFiles":[{"filename":"dist/bundle.js","reason":"builtin"}]}
+{"run":{"timestamp":"2026-02-18T14:32:15.123Z","durationMs":1000,"cwd":"/test","runId":"skip-123"},"type":"summary","totalFindings":0,"bySeverity":{"critical":0,"high":0,"medium":0,"low":0,"info":0},"totalSkippedFiles":1}
+`;
+
+    const result = parseJsonlReports(jsonlContent);
+
+    expect(result.reports[0]!.skippedFiles).toBeDefined();
+    expect(result.reports[0]!.skippedFiles!.length).toBe(1);
+    expect(result.reports[0]!.skippedFiles![0]!.filename).toBe('dist/bundle.js');
+    expect(result.reports[0]!.skippedFiles![0]!.reason).toBe('builtin');
+  });
+
+  it('round-trips through write and parse', () => {
+    const original: SkillReport[] = [
+      {
+        skill: 'test-skill',
+        summary: 'Found issues',
+        findings: [
+          { id: 'test-1', severity: 'high', title: 'Test Finding', description: 'Test description' },
+        ],
+        durationMs: 1500,
+        usage: { inputTokens: 500, outputTokens: 250, costUSD: 0.005 },
+        files: [
+          { filename: 'src/test.ts', findingCount: 1, durationMs: 1500 },
+        ],
+      },
+    ];
+
+    // Write to JSONL string
+    const jsonlContent = renderJsonlString(original, 2000, { runId: 'round-trip-123' });
+
+    // Parse back
+    const result = parseJsonlReports(jsonlContent);
+
+    expect(result.reports.length).toBe(1);
+    expect(result.reports[0]!.skill).toBe('test-skill');
+    expect(result.reports[0]!.summary).toBe('Found issues');
+    expect(result.reports[0]!.findings.length).toBe(1);
+    expect(result.reports[0]!.findings[0]!.id).toBe('test-1');
+    expect(result.reports[0]!.durationMs).toBe(1500);
+    expect(result.reports[0]!.usage?.inputTokens).toBe(500);
+    expect(result.reports[0]!.files?.length).toBe(1);
+    expect(result.reports[0]!.files![0]!.findingCount).toBe(1);
+  });
+});