feat: add npx CLI with monorepo, CI/CD, and ephemeral worker architecture (#256)

* feat: integrate npx CLI, CI/CD, and ephemeral worker architecture Bring in changes from shannon-npx: npx-distributable CLI package (cli/), semantic-release CI/CD workflows, ephemeral per-scan worker containers, TOML config support, setup wizard, and workspace management. Preserves all shannon-only changes: security hardening (localhost-bound ports, MCP env allowlist, path traversal guard), updated benchmarks (XBEN 19/31/35/44), README assets, and prompt injection disclaimer. Applies security hardening to cli/infra/compose.yml as well. * refactor: migrate to Turborepo + pnpm + Biome monorepo Restructure into apps/worker, apps/cli, packages/mcp-server with Turborepo task orchestration, pnpm workspaces, Biome linting/formatting, and tsdown CLI bundling. Key changes: - src/ -> apps/worker/src/, cli/ -> apps/cli/, mcp-server/ -> packages/mcp-server/ - prompts/ and configs/ moved into apps/worker/ - npm replaced with pnpm, package-lock.json replaced with pnpm-lock.yaml - Dockerfile updated for pnpm-based builds - CLI logs command rewritten with chokidar for cross-platform reliability - Router health checking added for auto-detected router mode - Centralized path resolution via apps/worker/src/paths.ts * fix: resolve all biome warnings and formatting issues - Remove unnecessary non-null assertions where values are guaranteed - Replace array index access with .at() for safer element retrieval - Use local variables to avoid repeated process.env lookups - Replace any types with unknown in functional utilities - Use nullish coalescing for TOTP hash byte access - Auto-format security patches to match biome config * fix: pin pnpm to 10.12.1 in Dockerfile for catalog support * fix: handle Esc cancellation in Bedrock setup flow Replace p.group() with individual prompts and per-field cancel checks, matching the pattern used by all other provider setup flows. * feat: add optional model customization to Anthropic setup * fix: resolve Docker bind mount permission errors on Linux Use entrypoint-based UID remapping instead of --user flag so the container's pentest user matches the host UID/GID, keeping bind-mounted volumes writable. Git config moved to --system level to survive remapping. * fix: show resumed workflow ID in splash screen URL When resuming a workflow, the Temporal Web UI link pointed to the old (terminated) workflow ID. Now extracts "New Workflow ID" from the resume header in workflow.log, falling back to the original ID for fresh scans. * style: fix biome formatting in docker.ts * fix: align TypeScript config types with JSON Schema - SuccessCondition.type: use schema values (url_contains, element_present, url_equals_exactly, text_contains) instead of stale values (url, cookie, element, redirect) - Authentication.login_flow: mark optional to match schema which does not require it * feat: mark GitHub release as latest during rollback * fix: use native ARM64 runners for Docker multi-platform builds Replace QEMU emulation with parallel native builds using a matrix strategy (ubuntu-latest for amd64, ubuntu-24.04-arm for arm64). Each platform pushes by digest, then a merge job creates the multi-arch manifest list before signing with cosign. * fix: resolve SessionMutex race condition with 3+ concurrent waiters * fix: skip POSIX permission check on Windows writeFileSync mode option is ignored on Windows, so config.toml gets 0o666 and the guard rejects it. * fix: resolve unsubstituted placeholders in report prompt Remove unused {{GITHUB_URL}} placeholder and wire up {{AUTH_CONTEXT}} with structured auth context (login type, username, URL, MFA status). * fix: remove duplicate environment gate from merge-docker job Move DOCKERHUB_USERNAME from vars to secrets so merge-docker can access credentials without its own environment scope. This eliminates the redundant double approval since build-docker already gates on release-publish. * fix: replace POSIX sleep binary with cross-platform async sleep execFileSync('sleep') is unavailable on Windows. Use node:timers/promises setTimeout instead, making ensureInfra async. * fix: use session.json for workflow ID on resume instead of parsing workflow.log On resume, workflow.log already exists with stale headers from the previous run. The CLI poll found '====' immediately and extracted the old workflow ID, producing a wrong Temporal Web UI URL. Read the workflow ID from session.json instead — the worker writes resume attempts there atomically. For fresh runs, poll until originalWorkflowId appears. For resumes, poll until a new resumeAttempts entry is appended. * feat: add custom base URL support for Anthropic-compatible proxies Support ANTHROPIC_BASE_URL + ANTHROPIC_AUTH_TOKEN to route SDK requests through LiteLLM or any Anthropic-compatible proxy. Adds TUI wizard option, TOML config mapping, credential validation, and preflight endpoint reachability check via SDK query. * fix: remove environment gates and add NPM_TOKEN to publish step * feat: add beta release and rollback workflows with cosign signing * fix: remove redundant checkout and pnpm steps from beta release workflow * docs: normalize README commands to mode-neutral shorthand Add a substitution note after Quick Start sections so all subsequent examples use bare `shannon` instead of mixing `./shannon` and `npx @keygraph/shannon`. Mode-specific commands (build, update, uninstall) get inline annotations. Also fixes a broken command in the Custom Base URL section. * fix: remove redundant `update` command Image is already auto-pulled by `ensureImage()` during `start` when the pinned version tag is missing locally. Manual `update` was unnecessary. * docs: add CLI package README stub * docs: update README setup instructions for dual CLI modes * docs: update announcement banner to npx availability * feat: migrate from MCP tools to CLI based tools (#252) * feat: migrate from MCP tools to CLI tools * fix: restore browser action emoji formatters for CLI output Adapt formatBrowserAction for playwright-cli commands, replacing the old mcp__playwright__browser_* tool name matching removed during migration. * fix: mount credential file to fixed container path for Vertex AI GOOGLE_APPLICATION_CREDENTIALS was forwarded as-is to the container, causing the relative host path to resolve against the repo mount instead of the credentials mount. Now both local and npx modes mount the resolved file to /app/credentials/google-sa-key.json and rewrite the env var to match. * feat: add git awareness and optional description field to config * fix: drop redundant --ipc host flag from worker container * fix: align announcement banner URL with main branch * feat: add target URL reachability preflight check (#254) * Moving asset benchmark graph image to this folder * Move benchmark results to benchmark repo Windows Defender flags exploit code in the pentest reports as false positives, forcing every Windows user to add a Defender exclusion just to clone Shannon. * Updated README * fix: case-insensitive grep for semantic-release version probe * fix: harden supply chain security (#255) * fix: patch smol-toml and tsdown vulnerabilities Update smol-toml 1.6.0→1.6.1 (DoS via recursive comment parsing) and tsdown 0.21.2→0.21.5 (picomatch ReDoS + method injection). * fix: pin all unpinned dependency versions in Dockerfile Pins subfinder v2.13.0, WhatWeb v0.6.3 (switched from git clone to release tarball), schemathesis 4.13.0, addressable 2.8.9, claude-code 2.1.84, and playwright-cli 0.1.1 for reproducible builds. * fix: pin GitHub Actions to commit SHAs for supply chain security * fix: pin GitHub Actions to commit SHAs in beta and rollback workflows
2026-03-27 02:34:29 +05:30
parent 0d172f5e32
commit bc8fd203ed
4058 changed files with 7774 additions and 1189080 deletions
@@ -0,0 +1,646 @@
+// Copyright (C) 2025 Keygraph, Inc.
+//
+// This program is free software: you can redistribute it and/or modify
+// it under the terms of the GNU Affero General Public License version 3
+// as published by the Free Software Foundation.
+
+/**
+ * Temporal activities for Shannon agent execution.
+ *
+ * Each activity wraps service calls with Temporal-specific concerns:
+ * - Heartbeat loop (2s interval) to signal worker liveness
+ * - Error classification into ApplicationFailure
+ * - Container lifecycle management
+ *
+ * Business logic is delegated to services in src/services/.
+ */
+
+import fs from 'node:fs/promises';
+import path from 'node:path';
+import { ApplicationFailure, Context, heartbeat } from '@temporalio/activity';
+import { AuditSession } from '../audit/index.js';
+import type { ResumeAttempt } from '../audit/metrics-tracker.js';
+import { copyDeliverablesToAudit, type SessionMetadata } from '../audit/utils.js';
+import type { WorkflowSummary } from '../audit/workflow-logger.js';
+import { getContainer, getOrCreateContainer, removeContainer } from '../services/container.js';
+import { classifyErrorForTemporal, PentestError } from '../services/error-handling.js';
+import { ExploitationCheckerService } from '../services/exploitation-checker.js';
+import { executeGitCommandWithRetry } from '../services/git-manager.js';
+import { runPreflightChecks } from '../services/preflight.js';
+import type { ExploitationDecision, VulnType } from '../services/queue-validation.js';
+import { assembleFinalReport, injectModelIntoReport } from '../services/reporting.js';
+import { AGENTS } from '../session-manager.js';
+import type { AgentName } from '../types/agents.js';
+import { ALL_AGENTS } from '../types/agents.js';
+import { ErrorCode } from '../types/errors.js';
+import { isErr } from '../types/result.js';
+import { fileExists, readJson } from '../utils/file-io.js';
+import { createActivityLogger } from './activity-logger.js';
+import type { AgentMetrics, ResumeState } from './shared.js';
+
+// Max lengths to prevent Temporal protobuf buffer overflow
+const MAX_ERROR_MESSAGE_LENGTH = 2000;
+const MAX_STACK_TRACE_LENGTH = 1000;
+
+// Max retries for output validation errors (agent didn't save deliverables)
+const MAX_OUTPUT_VALIDATION_RETRIES = 3;
+
+const HEARTBEAT_INTERVAL_MS = 2000;
+
+/**
+ * Input for all agent activities.
+ */
+export interface ActivityInput {
+  webUrl: string;
+  repoPath: string;
+  configPath?: string;
+  outputPath?: string;
+  pipelineTestingMode?: boolean;
+  workflowId: string;
+  sessionId: string;
+}
+
+/**
+ * Truncate error message to prevent buffer overflow in Temporal serialization.
+ */
+function truncateErrorMessage(message: string): string {
+  if (message.length <= MAX_ERROR_MESSAGE_LENGTH) {
+    return message;
+  }
+  return `${message.slice(0, MAX_ERROR_MESSAGE_LENGTH - 20)}\n[truncated]`;
+}
+
+/**
+ * Truncate stack trace on an ApplicationFailure to prevent buffer overflow.
+ */
+function truncateStackTrace(failure: ApplicationFailure): void {
+  if (failure.stack && failure.stack.length > MAX_STACK_TRACE_LENGTH) {
+    failure.stack = `${failure.stack.slice(0, MAX_STACK_TRACE_LENGTH)}\n[stack truncated]`;
+  }
+}
+
+/**
+ * Build SessionMetadata from ActivityInput.
+ */
+function buildSessionMetadata(input: ActivityInput): SessionMetadata {
+  const { webUrl, repoPath, outputPath, sessionId } = input;
+  return {
+    id: sessionId,
+    webUrl,
+    repoPath,
+    ...(outputPath && { outputPath }),
+  };
+}
+
+/**
+ * Core activity implementation using services.
+ *
+ * Executes a single agent with:
+ * 1. Heartbeat loop for worker liveness
+ * 2. Container creation/reuse
+ * 3. Service-based agent execution
+ * 4. Error classification for Temporal retry
+ */
+async function runAgentActivity(agentName: AgentName, input: ActivityInput): Promise<AgentMetrics> {
+  const { repoPath, configPath, pipelineTestingMode = false, workflowId, webUrl } = input;
+  const startTime = Date.now();
+  const attemptNumber = Context.current().info.attempt;
+
+  // Heartbeat loop - signals worker is alive to Temporal server
+  const heartbeatInterval = setInterval(() => {
+    const elapsed = Math.floor((Date.now() - startTime) / 1000);
+    heartbeat({ agent: agentName, elapsedSeconds: elapsed, attempt: attemptNumber });
+  }, HEARTBEAT_INTERVAL_MS);
+
+  try {
+    const logger = createActivityLogger();
+
+    // 1. Build session metadata and get/create container
+    const sessionMetadata = buildSessionMetadata(input);
+    const container = getOrCreateContainer(workflowId, sessionMetadata);
+
+    // 2. Create audit session for THIS agent execution
+    // NOTE: Each agent needs its own AuditSession because AuditSession uses
+    // instance state (currentAgentName) that cannot be shared across parallel agents
+    const auditSession = new AuditSession(sessionMetadata);
+    await auditSession.initialize(workflowId);
+
+    // 3. Execute agent via service (throws PentestError on failure)
+    const endResult = await container.agentExecution.executeOrThrow(
+      agentName,
+      {
+        webUrl,
+        repoPath,
+        configPath,
+        pipelineTestingMode,
+        attemptNumber,
+      },
+      auditSession,
+      logger,
+    );
+
+    // 4. Return metrics
+    return {
+      durationMs: Date.now() - startTime,
+      inputTokens: null,
+      outputTokens: null,
+      costUsd: endResult.cost_usd,
+      numTurns: null,
+      model: endResult.model,
+    };
+  } catch (error) {
+    // If error is already an ApplicationFailure, re-throw directly
+    if (error instanceof ApplicationFailure) {
+      throw error;
+    }
+
+    // Check if output validation retry limit reached (PentestError with code)
+    if (
+      error instanceof PentestError &&
+      error.code === ErrorCode.OUTPUT_VALIDATION_FAILED &&
+      attemptNumber >= MAX_OUTPUT_VALIDATION_RETRIES
+    ) {
+      throw ApplicationFailure.nonRetryable(
+        `Agent ${agentName} failed output validation after ${attemptNumber} attempts`,
+        'OutputValidationError',
+        [{ agentName, attemptNumber, elapsed: Date.now() - startTime }],
+      );
+    }
+
+    // Classify error for Temporal retry behavior
+    const classified = classifyErrorForTemporal(error);
+    const rawMessage = error instanceof Error ? error.message : String(error);
+    const message = truncateErrorMessage(rawMessage);
+
+    if (classified.retryable) {
+      const failure = ApplicationFailure.create({
+        message,
+        type: classified.type,
+        details: [{ agentName, attemptNumber, elapsed: Date.now() - startTime }],
+      });
+      truncateStackTrace(failure);
+      throw failure;
+    } else {
+      const failure = ApplicationFailure.nonRetryable(message, classified.type, [
+        { agentName, attemptNumber, elapsed: Date.now() - startTime },
+      ]);
+      truncateStackTrace(failure);
+      throw failure;
+    }
+  } finally {
+    clearInterval(heartbeatInterval);
+  }
+}
+
+export async function runPreReconAgent(input: ActivityInput): Promise<AgentMetrics> {
+  return runAgentActivity('pre-recon', input);
+}
+
+export async function runReconAgent(input: ActivityInput): Promise<AgentMetrics> {
+  return runAgentActivity('recon', input);
+}
+
+export async function runInjectionVulnAgent(input: ActivityInput): Promise<AgentMetrics> {
+  return runAgentActivity('injection-vuln', input);
+}
+
+export async function runXssVulnAgent(input: ActivityInput): Promise<AgentMetrics> {
+  return runAgentActivity('xss-vuln', input);
+}
+
+export async function runAuthVulnAgent(input: ActivityInput): Promise<AgentMetrics> {
+  return runAgentActivity('auth-vuln', input);
+}
+
+export async function runSsrfVulnAgent(input: ActivityInput): Promise<AgentMetrics> {
+  return runAgentActivity('ssrf-vuln', input);
+}
+
+export async function runAuthzVulnAgent(input: ActivityInput): Promise<AgentMetrics> {
+  return runAgentActivity('authz-vuln', input);
+}
+
+export async function runInjectionExploitAgent(input: ActivityInput): Promise<AgentMetrics> {
+  return runAgentActivity('injection-exploit', input);
+}
+
+export async function runXssExploitAgent(input: ActivityInput): Promise<AgentMetrics> {
+  return runAgentActivity('xss-exploit', input);
+}
+
+export async function runAuthExploitAgent(input: ActivityInput): Promise<AgentMetrics> {
+  return runAgentActivity('auth-exploit', input);
+}
+
+export async function runSsrfExploitAgent(input: ActivityInput): Promise<AgentMetrics> {
+  return runAgentActivity('ssrf-exploit', input);
+}
+
+export async function runAuthzExploitAgent(input: ActivityInput): Promise<AgentMetrics> {
+  return runAgentActivity('authz-exploit', input);
+}
+
+export async function runReportAgent(input: ActivityInput): Promise<AgentMetrics> {
+  return runAgentActivity('report', input);
+}
+
+/**
+ * Preflight validation activity.
+ *
+ * Runs cheap checks before any agent execution:
+ * 1. Repository path exists with .git
+ * 2. Config file validates (if provided)
+ * 3. Credential validation (API key, OAuth, or router mode)
+ * 4. Target URL reachable from the container
+ *
+ * NOT using runAgentActivity — preflight doesn't run an agent via the SDK.
+ */
+export async function runPreflightValidation(input: ActivityInput): Promise<void> {
+  const startTime = Date.now();
+  const attemptNumber = Context.current().info.attempt;
+
+  const heartbeatInterval = setInterval(() => {
+    const elapsed = Math.floor((Date.now() - startTime) / 1000);
+    heartbeat({ phase: 'preflight', elapsedSeconds: elapsed, attempt: attemptNumber });
+  }, HEARTBEAT_INTERVAL_MS);
+
+  try {
+    const logger = createActivityLogger();
+    logger.info('Running preflight validation...', { attempt: attemptNumber });
+
+    const result = await runPreflightChecks(input.webUrl, input.repoPath, input.configPath, logger);
+
+    if (isErr(result)) {
+      const classified = classifyErrorForTemporal(result.error);
+      const message = truncateErrorMessage(result.error.message);
+
+      if (classified.retryable) {
+        const failure = ApplicationFailure.create({
+          message,
+          type: classified.type,
+          details: [{ phase: 'preflight', attemptNumber, elapsed: Date.now() - startTime }],
+        });
+        truncateStackTrace(failure);
+        throw failure;
+      } else {
+        const failure = ApplicationFailure.nonRetryable(message, classified.type, [
+          { phase: 'preflight', attemptNumber, elapsed: Date.now() - startTime },
+        ]);
+        truncateStackTrace(failure);
+        throw failure;
+      }
+    }
+
+    logger.info('Preflight validation passed');
+  } catch (error) {
+    if (error instanceof ApplicationFailure) {
+      throw error;
+    }
+
+    const classified = classifyErrorForTemporal(error);
+    const rawMessage = error instanceof Error ? error.message : String(error);
+    const message = truncateErrorMessage(rawMessage);
+
+    const failure = ApplicationFailure.nonRetryable(message, classified.type, [
+      { phase: 'preflight', attemptNumber, elapsed: Date.now() - startTime },
+    ]);
+    truncateStackTrace(failure);
+    throw failure;
+  } finally {
+    clearInterval(heartbeatInterval);
+  }
+}
+
+/**
+ * Assemble the final report by concatenating exploitation evidence files.
+ */
+export async function assembleReportActivity(input: ActivityInput): Promise<void> {
+  const { repoPath } = input;
+  const logger = createActivityLogger();
+  logger.info('Assembling deliverables from specialist agents...');
+  try {
+    await assembleFinalReport(repoPath, logger);
+  } catch (error) {
+    const err = error as Error;
+    logger.warn(`Error assembling final report: ${err.message}`);
+  }
+}
+
+/**
+ * Inject model metadata into the final report.
+ */
+export async function injectReportMetadataActivity(input: ActivityInput): Promise<void> {
+  const { repoPath, sessionId, outputPath } = input;
+  const logger = createActivityLogger();
+  const effectiveOutputPath = outputPath ? path.join(outputPath, sessionId) : path.join('./workspaces', sessionId);
+  try {
+    await injectModelIntoReport(repoPath, effectiveOutputPath, logger);
+  } catch (error) {
+    const err = error as Error;
+    logger.warn(`Error injecting model into report: ${err.message}`);
+  }
+}
+
+/**
+ * Check if exploitation should run for a given vulnerability type.
+ *
+ * Uses existing container if available (from prior agent runs),
+ * otherwise creates service directly (stateless, no dependencies).
+ */
+export async function checkExploitationQueue(input: ActivityInput, vulnType: VulnType): Promise<ExploitationDecision> {
+  const { repoPath, workflowId } = input;
+  const logger = createActivityLogger();
+
+  // Reuse container's service if available (from prior vuln agent runs)
+  const existingContainer = getContainer(workflowId);
+  const checker = existingContainer?.exploitationChecker ?? new ExploitationCheckerService();
+
+  return checker.checkQueue(vulnType, repoPath, logger);
+}
+
+interface SessionJson {
+  session: {
+    id: string;
+    webUrl: string;
+    repoPath?: string;
+    originalWorkflowId?: string;
+    resumeAttempts?: ResumeAttempt[];
+  };
+  metrics: {
+    agents: Record<
+      string,
+      {
+        status: 'in-progress' | 'success' | 'failed';
+        checkpoint?: string;
+      }
+    >;
+  };
+}
+
+/**
+ * Load resume state from an existing workspace.
+ */
+export async function loadResumeState(
+  workspaceName: string,
+  expectedUrl: string,
+  expectedRepoPath: string,
+): Promise<ResumeState> {
+  // 1. Validate workspace exists
+  const sessionPath = path.join('./workspaces', workspaceName, 'session.json');
+
+  const exists = await fileExists(sessionPath);
+  if (!exists) {
+    throw ApplicationFailure.nonRetryable(
+      `Workspace not found: ${workspaceName}\nExpected path: ${sessionPath}`,
+      'WorkspaceNotFoundError',
+    );
+  }
+
+  // 2. Parse session.json and validate URL match
+  let session: SessionJson;
+  try {
+    session = await readJson<SessionJson>(sessionPath);
+  } catch (error) {
+    const errorMsg = error instanceof Error ? error.message : String(error);
+    throw ApplicationFailure.nonRetryable(
+      `Corrupted session.json in workspace ${workspaceName}: ${errorMsg}`,
+      'CorruptedSessionError',
+    );
+  }
+
+  if (session.session.webUrl !== expectedUrl) {
+    throw ApplicationFailure.nonRetryable(
+      `URL mismatch with workspace\n  Workspace URL: ${session.session.webUrl}\n  Provided URL:  ${expectedUrl}`,
+      'URLMismatchError',
+    );
+  }
+
+  // 3. Cross-check agent status with deliverables on disk
+  const completedAgents: string[] = [];
+  const agents = session.metrics.agents;
+
+  for (const agentName of ALL_AGENTS) {
+    const agentData = agents[agentName];
+    if (!agentData || agentData.status !== 'success') {
+      continue;
+    }
+
+    const deliverableFilename = AGENTS[agentName].deliverableFilename;
+    const deliverablePath = `${expectedRepoPath}/deliverables/${deliverableFilename}`;
+    const deliverableExists = await fileExists(deliverablePath);
+
+    if (!deliverableExists) {
+      const logger = createActivityLogger();
+      logger.warn(`Agent ${agentName} shows success but deliverable missing, will re-run`);
+      continue;
+    }
+
+    completedAgents.push(agentName);
+  }
+
+  // 4. Collect git checkpoints and validate at least one exists
+  const checkpoints = completedAgents
+    .map((name) => agents[name]?.checkpoint)
+    .filter((hash): hash is string => hash != null);
+
+  if (checkpoints.length === 0) {
+    const successAgents = Object.entries(agents)
+      .filter(([, data]) => data.status === 'success')
+      .map(([name]) => name);
+
+    throw ApplicationFailure.nonRetryable(
+      `Cannot resume workspace ${workspaceName}: ` +
+        (successAgents.length > 0
+          ? `${successAgents.length} agent(s) show success in session.json (${successAgents.join(', ')}) ` +
+            `but their deliverable files are missing from disk. ` +
+            `Start a fresh run instead.`
+          : `No agents completed successfully. Start a fresh run instead.`),
+      'NoCheckpointsError',
+    );
+  }
+
+  // 5. Find the most recent checkpoint commit
+  const checkpointHash = await findLatestCommit(expectedRepoPath, checkpoints);
+  const originalWorkflowId = session.session.originalWorkflowId || session.session.id;
+
+  // 6. Log summary and return resume state
+  const logger = createActivityLogger();
+  logger.info('Resume state loaded', {
+    workspace: workspaceName,
+    completedAgents: completedAgents.length,
+    checkpoint: checkpointHash,
+  });
+
+  return {
+    workspaceName,
+    originalUrl: session.session.webUrl,
+    completedAgents,
+    checkpointHash,
+    originalWorkflowId,
+  };
+}
+
+async function findLatestCommit(repoPath: string, commitHashes: string[]): Promise<string> {
+  if (commitHashes.length === 1) {
+    const hash = commitHashes[0];
+    if (!hash) {
+      throw new PentestError(
+        'Empty commit hash in array',
+        'filesystem',
+        false, // Non-retryable - corrupt workspace state
+        { phase: 'resume' },
+        ErrorCode.GIT_CHECKPOINT_FAILED,
+      );
+    }
+    return hash;
+  }
+
+  const result = await executeGitCommandWithRetry(
+    ['git', 'rev-list', '--max-count=1', ...commitHashes],
+    repoPath,
+    'find latest commit',
+  );
+
+  return result.stdout.trim();
+}
+
+/**
+ * Restore git workspace to a checkpoint and clean up partial deliverables.
+ */
+export async function restoreGitCheckpoint(
+  repoPath: string,
+  checkpointHash: string,
+  incompleteAgents: AgentName[],
+): Promise<void> {
+  const logger = createActivityLogger();
+  logger.info(`Restoring git workspace to ${checkpointHash}...`);
+
+  await executeGitCommandWithRetry(
+    ['git', 'reset', '--hard', checkpointHash],
+    repoPath,
+    'reset to checkpoint for resume',
+  );
+  await executeGitCommandWithRetry(['git', 'clean', '-fd'], repoPath, 'clean untracked files for resume');
+
+  for (const agentName of incompleteAgents) {
+    const deliverableFilename = AGENTS[agentName].deliverableFilename;
+    const deliverablePath = `${repoPath}/deliverables/${deliverableFilename}`;
+    try {
+      const exists = await fileExists(deliverablePath);
+      if (exists) {
+        logger.warn(`Cleaning partial deliverable: ${agentName}`);
+        await fs.unlink(deliverablePath);
+      }
+    } catch (error) {
+      logger.info(`Note: Failed to delete ${deliverablePath}: ${error}`);
+    }
+  }
+
+  logger.info('Workspace restored to clean state');
+}
+
+/**
+ * Record a resume attempt in session.json and write resume header to workflow.log.
+ */
+export async function recordResumeAttempt(
+  input: ActivityInput,
+  terminatedWorkflows: string[],
+  checkpointHash: string,
+  previousWorkflowId: string,
+  completedAgents: string[],
+): Promise<void> {
+  const sessionMetadata = buildSessionMetadata(input);
+  const auditSession = new AuditSession(sessionMetadata);
+  await auditSession.initialize();
+
+  // Update session.json with resume attempt
+  await auditSession.addResumeAttempt(input.workflowId, terminatedWorkflows, checkpointHash);
+
+  // Write resume header to workflow.log
+  await auditSession.logResumeHeader({
+    previousWorkflowId,
+    newWorkflowId: input.workflowId,
+    checkpointHash,
+    completedAgents,
+  });
+}
+
+/**
+ * Log phase transition to the unified workflow log.
+ */
+export async function logPhaseTransition(
+  input: ActivityInput,
+  phase: string,
+  event: 'start' | 'complete',
+): Promise<void> {
+  const sessionMetadata = buildSessionMetadata(input);
+  const auditSession = new AuditSession(sessionMetadata);
+  await auditSession.initialize(input.workflowId);
+
+  if (event === 'start') {
+    await auditSession.logPhaseStart(phase);
+  } else {
+    await auditSession.logPhaseComplete(phase);
+  }
+}
+
+/**
+ * Log workflow completion with full summary.
+ * Cleans up container when done.
+ */
+export async function logWorkflowComplete(input: ActivityInput, summary: WorkflowSummary): Promise<void> {
+  const { repoPath, workflowId } = input;
+  const sessionMetadata = buildSessionMetadata(input);
+
+  // 1. Initialize audit session and mark final status
+  const auditSession = new AuditSession(sessionMetadata);
+  await auditSession.initialize(workflowId);
+  await auditSession.updateSessionStatus(summary.status);
+
+  // 2. Load cumulative metrics from session.json
+  const sessionData = (await auditSession.getMetrics()) as {
+    metrics: {
+      total_duration_ms: number;
+      total_cost_usd: number;
+      agents: Record<string, { final_duration_ms: number; total_cost_usd: number }>;
+    };
+  };
+
+  // 3. Fill in metrics for skipped agents (resumed from previous run)
+  const agentMetrics = { ...summary.agentMetrics };
+  for (const agentName of summary.completedAgents) {
+    if (!agentMetrics[agentName]) {
+      const agentData = sessionData.metrics.agents[agentName];
+      if (agentData) {
+        agentMetrics[agentName] = {
+          durationMs: agentData.final_duration_ms,
+          costUsd: agentData.total_cost_usd,
+        };
+      }
+    }
+  }
+
+  // 4. Build cumulative summary with cross-run totals
+  const cumulativeSummary: WorkflowSummary = {
+    ...summary,
+    totalDurationMs: sessionData.metrics.total_duration_ms,
+    totalCostUsd: sessionData.metrics.total_cost_usd,
+    agentMetrics,
+  };
+
+  // 5. Write completion entry to workflow.log
+  await auditSession.logWorkflowComplete(cumulativeSummary);
+
+  // 6. Copy deliverables to workspaces
+  try {
+    await copyDeliverablesToAudit(sessionMetadata, repoPath);
+  } catch (copyErr) {
+    const logger = createActivityLogger();
+    logger.error('Failed to copy deliverables to workspaces', {
+      error: copyErr instanceof Error ? copyErr.message : String(copyErr),
+    });
+  }
+
+  // 7. Clean up container
+  removeContainer(workflowId);
+}