Pre-launch — Gluecron is in final validation. Public signups and git hosting for non-owner users open after launch review.
CodeIssuesDiscussionsWikiPull RequestsProjectsCommitsActionsReleasesContributorsPulse● GatesSecuritySettingsDeploymentsPipelineInsightsAgents✨ Explain✨ Ask AI✨ Workspace✨ Spec✨ Tests▓ Debt Map✨ NL Search🏛 Archaeology
claude/adoring-hopper-5x74bqclaude/affectionate-feynman-ykrf1hclaude/architecture-audit-design-wxprenclaude/build-status-update-3MXsfclaude/charming-meitner-mllb5rclaude/compare-gate-gluecron-s4mFQclaude/confident-faraday-tikcwbclaude/continue-work-XMTlIclaude/crontech-gluecron-deploy-7MIECclaude/crontech-platform-setup-SeKfwclaude/design-2026claude/ecstatic-ptolemy-jMdigclaude/enhance-github-integration-QNHdGclaude/fix-aa-loop-issue-PonMQclaude/fix-actions-and-processclaude/fix-desktop-errors-XqoW8claude/fix-red-workflowsclaude/fix-website-access-6FKJNclaude/gatetest-integration-hardeningclaude/github-audit-improvements-bDFr9claude/gluecron-launch-status-FoMRlclaude/hopeful-lamport-olfCTclaude/issue-to-pr-and-protectionsclaude/jolly-heisenberg-2sg1Qclaude/launch-preparation-QmTb6claude/new-session-xk1l7claude/plan-platform-architecture-kkN4yclaude/platform-analysis-roadmap-1nUGLclaude/platform-launch-assessment-8dWV8claude/polish-platform-release-AeDrUclaude/resume-previous-work-KzyLwclaude/review-crontech-handoff-qYEVqclaude/review-project-completeness-lHhS2claude/review-readme-docs-ulqPKclaude/serene-edison-rj87weclaude/setup-multi-repo-dev-BCwNQclaude/ship-fixes-and-tests-Jvz1cclaude/site-audit-competitive-pctlwgclaude/site-migration-vercel-XstpKclaude/standalone-product-repos-XHFTDcopilot/feat-smart-empty-states-keyboard-first-enhancementcopilot/feat-smart-morning-digest-review-context-restorecopilot/fix-and-process-workflowscopilot/update-ai-powered-code-reviewfeat/debt-mapfeat/push-policy-codeowners-hardeningfeat/smart-digest-contextfeat/stage-impactfeat/t1-secret-migrationfeat/u-polishfeat/w-self-hostfeat/w2-claude-configfix/agent-journey-orphan-sweepgatetest/auto-fix-1776586424172gatetest/auto-fix-1776586534814gatetest/auto-fix-1776590685143gatetest/auto-fix-1776590808199mainops/redeploy-retriggerstyle/dxt-cta-themeworktree-agent-a3377aad30d55da26worktree-agent-a7ef607b7ee1d6c74
ai-workspace.ts16.6 KB · 587 lines
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
/**
 * AI Copilot Workspace — issue-to-PR autonomous agent.
 *
 * Pipeline:
 *   1. Load issue context (title + body + recent comments).
 *   2. Explore the codebase — ls-tree then ask Claude which files to read.
 *   3. Generate an implementation plan (Claude call) + post as issue comment.
 *   4. Implement: create branch, apply file edits via git plumbing.
 *   5. Open a draft PR linked to the issue.
 *
 * The hot path is fully in-memory (WorkspaceJob map). A workspace_jobs table
 * is written for auditability via `src/db/schema.ts` + migration 0103.
 *
 * Never throws externally. All failures set job.status = "failed".
 */

import { join } from "path";
import { eq, desc } from "drizzle-orm";
import { db } from "../db";
import {
  issues,
  issueComments,
  repositories,
  users,
  pullRequests,
  workspaceJobs,
} from "../db/schema";
import { getAnthropic, isAiAvailable, MODEL_SONNET, extractText, parseJsonResponse } from "./ai-client";
import { applyEditsToNewBranch, type FileEdit } from "./spec-git";
import { getBlob } from "../git/repository";
import { getBotUserIdOrFallback } from "./bot-user";
import { config } from "./config";

// ---------------------------------------------------------------------------
// Public types
// ---------------------------------------------------------------------------

export type WorkspaceStatus =
  | "pending"
  | "planning"
  | "implementing"
  | "opening_pr"
  | "done"
  | "failed";

export interface WorkspaceJob {
  id: string;
  repoId: string;
  issueId: string;
  issueNumber: number;
  ownerName: string;
  repoName: string;
  status: WorkspaceStatus;
  planComment?: string;   // the plan posted as issue comment
  branchName?: string;    // e.g. "workspace/issue-42-fix-auth"
  prNumber?: number;
  errorMessage?: string;
  startedAt: Date;
  updatedAt: Date;
}

// ---------------------------------------------------------------------------
// In-memory job store (LRU, max 100 jobs)
// ---------------------------------------------------------------------------

const MAX_JOBS = 100;
const jobById = new Map<string, WorkspaceJob>();
// issueId → jobId for O(1) lookup by issue
const jobByIssueId = new Map<string, string>();
// repoId → jobId for "one active job per repo" guard
const activeJobByRepoId = new Map<string, string>();

function evictOldestIfNeeded(): void {
  if (jobById.size < MAX_JOBS) return;
  // Evict oldest (first inserted) entry
  const firstKey = jobById.keys().next().value;
  if (firstKey) {
    const old = jobById.get(firstKey);
    if (old) {
      jobByIssueId.delete(old.issueId);
      if (activeJobByRepoId.get(old.repoId) === firstKey) {
        activeJobByRepoId.delete(old.repoId);
      }
    }
    jobById.delete(firstKey);
  }
}

function storeJob(job: WorkspaceJob): void {
  evictOldestIfNeeded();
  jobById.set(job.id, job);
  jobByIssueId.set(job.issueId, job.id);
  if (job.status !== "done" && job.status !== "failed") {
    activeJobByRepoId.set(job.repoId, job.id);
  }
}

function updateJob(job: WorkspaceJob, patch: Partial<WorkspaceJob>): void {
  Object.assign(job, patch, { updatedAt: new Date() });
  if (job.status === "done" || job.status === "failed") {
    if (activeJobByRepoId.get(job.repoId) === job.id) {
      activeJobByRepoId.delete(job.repoId);
    }
  }
}

// ---------------------------------------------------------------------------
// Public accessors
// ---------------------------------------------------------------------------

export function getWorkspaceJob(jobId: string): WorkspaceJob | undefined {
  return jobById.get(jobId);
}

export function getWorkspaceJobForIssue(issueId: string): WorkspaceJob | undefined {
  const jobId = jobByIssueId.get(issueId);
  return jobId ? jobById.get(jobId) : undefined;
}

// ---------------------------------------------------------------------------
// Entry point
// ---------------------------------------------------------------------------

export async function startWorkspace(
  issueId: string,
  issueNumber: number,
  repoId: string,
  ownerName: string,
  repoName: string,
  triggeredByUserId: string
): Promise<WorkspaceJob> {
  // Guard: AI must be available
  if (!isAiAvailable()) {
    const failed: WorkspaceJob = {
      id: crypto.randomUUID(),
      repoId,
      issueId,
      issueNumber,
      ownerName,
      repoName,
      status: "failed",
      errorMessage: "ANTHROPIC_API_KEY is not configured",
      startedAt: new Date(),
      updatedAt: new Date(),
    };
    storeJob(failed);
    return failed;
  }

  // Guard: max 1 active job per repo
  const existingJobId = activeJobByRepoId.get(repoId);
  if (existingJobId) {
    const existing = jobById.get(existingJobId);
    if (existing && existing.status !== "done" && existing.status !== "failed") {
      return existing;
    }
    // Stale entry — clear it
    activeJobByRepoId.delete(repoId);
  }

  const job: WorkspaceJob = {
    id: crypto.randomUUID(),
    repoId,
    issueId,
    issueNumber,
    ownerName,
    repoName,
    status: "pending",
    startedAt: new Date(),
    updatedAt: new Date(),
  };

  storeJob(job);

  // Persist to DB for auditability (best-effort)
  try {
    await db.insert(workspaceJobs).values({
      id: job.id,
      repoId,
      issueId,
      triggeredBy: triggeredByUserId,
      status: "pending",
    });
  } catch {
    /* non-fatal — in-memory store is the source of truth */
  }

  // Fire-and-forget pipeline
  runWorkspacePipeline(job, triggeredByUserId).catch(() => {
    /* errors are captured inside the pipeline */
  });

  return job;
}

// ---------------------------------------------------------------------------
// Pipeline
// ---------------------------------------------------------------------------

async function runWorkspacePipeline(
  job: WorkspaceJob,
  triggeredByUserId: string
): Promise<void> {
  try {
    // Step 1 — Load issue context
    const issueCtx = await loadIssueContext(job.issueId);
    if (!issueCtx) {
      throw new Error("Issue not found or DB read failed");
    }

    // Step 2 — Explore codebase
    updateJob(job, { status: "planning" });
    await persistJobStatus(job);

    const reposBase = config.gitReposPath;
    const repoDiskPath = join(reposBase, job.ownerName, `${job.repoName}.git`);

    const fileTree = await getFileTree(repoDiskPath);
    const relevantFiles = await pickRelevantFiles(issueCtx, fileTree);
    const fileContents = await readFileContents(
      job.ownerName,
      job.repoName,
      relevantFiles
    );

    // Step 3 — Generate implementation plan
    const plan = await generatePlan(issueCtx, fileContents);

    // Post plan as issue comment
    const planBody = formatPlanComment(plan);
    const botUserId = await getBotUserIdOrFallback(triggeredByUserId);
    const commentId = await postIssueComment(job.issueId, botUserId, planBody);
    updateJob(job, { planComment: commentId ?? undefined });

    // Step 4 — Implement
    updateJob(job, { status: "implementing" });
    await persistJobStatus(job);

    // Resolve default branch
    const repoRow = await db
      .select({ defaultBranch: repositories.defaultBranch })
      .from(repositories)
      .where(eq(repositories.id, job.repoId))
      .limit(1);
    const baseBranch = repoRow[0]?.defaultBranch ?? "main";

    const branchName = plan.branchName || sanitizeBranchName(
      `workspace/issue-${job.issueNumber}-${slugify(issueCtx.title)}`
    );
    updateJob(job, { branchName });

    const edits = buildEdits(plan);
    if (edits.length === 0) {
      throw new Error("AI plan produced no file changes");
    }

    const commitMsg =
      `feat: implement #${job.issueNumber}${plan.summary.slice(0, 60)}\n\n` +
      `Closes #${job.issueNumber}\n\n` +
      `Generated by Gluecron AI Workspace.`;

    const applied = await applyEditsToNewBranch({
      repoDiskPath,
      baseRef: baseBranch,
      edits,
      branchName,
      commitMessage: commitMsg,
      authorName: "Gluecron Workspace",
      authorEmail: "workspace@gluecron.com",
    });

    if (!applied.ok) {
      throw new Error(`git apply failed: ${applied.error}`);
    }

    // Step 5 — Open draft PR
    updateJob(job, { status: "opening_pr" });
    await persistJobStatus(job);

    const prBody =
      `Closes #${job.issueNumber}\n\n` +
      `${plan.summary}\n\n` +
      `**Files changed:**\n` +
      applied.filesChanged.map((f) => `- \`${f}\``).join("\n") +
      `\n\n<!-- gluecron:workspace:pr -->`;

    const prTitle = `feat: ${plan.summary.slice(0, 140)}`;

    const [pr] = await db
      .insert(pullRequests)
      .values({
        repositoryId: job.repoId,
        authorId: botUserId,
        title: prTitle.slice(0, 200),
        body: prBody,
        baseBranch,
        headBranch: applied.branchName,
        isDraft: true,
      })
      .returning({ number: pullRequests.number });

    if (!pr?.number) {
      throw new Error("PR insert returned no number");
    }

    updateJob(job, { status: "done", prNumber: pr.number });
    await persistJobStatus(job);
  } catch (err: unknown) {
    const msg = err instanceof Error ? err.message : String(err);
    updateJob(job, { status: "failed", errorMessage: msg });
    await persistJobStatus(job);
  }
}

// ---------------------------------------------------------------------------
// Step helpers
// ---------------------------------------------------------------------------

interface IssueContext {
  title: string;
  body: string;
  comments: string[];
}

async function loadIssueContext(issueId: string): Promise<IssueContext | null> {
  try {
    const issueRows = await db
      .select({ title: issues.title, body: issues.body })
      .from(issues)
      .where(eq(issues.id, issueId))
      .limit(1);

    if (!issueRows[0]) return null;
    const { title, body } = issueRows[0];

    const commentRows = await db
      .select({ body: issueComments.body })
      .from(issueComments)
      .where(eq(issueComments.issueId, issueId))
      .orderBy(desc(issueComments.createdAt))
      .limit(5);

    return {
      title,
      body: body ?? "",
      comments: commentRows.map((r) => r.body),
    };
  } catch {
    return null;
  }
}

async function getFileTree(repoDiskPath: string): Promise<string[]> {
  try {
    const proc = Bun.spawn(
      ["git", "ls-tree", "-r", "--name-only", "HEAD"],
      { cwd: repoDiskPath, stdout: "pipe", stderr: "pipe" }
    );
    const text = await new Response(proc.stdout).text();
    await proc.exited;
    const lines = text
      .split("\n")
      .map((l) => l.trim())
      .filter(Boolean);
    return lines.slice(0, 500);
  } catch {
    return [];
  }
}

async function pickRelevantFiles(
  ctx: IssueContext,
  fileTree: string[]
): Promise<string[]> {
  if (fileTree.length === 0) return [];

  const prompt =
    `Given this issue:\n${ctx.title}\n${ctx.body}\n\n` +
    `And this file tree:\n${fileTree.join("\n")}\n\n` +
    `List the 20 most relevant file paths to read in order to implement the issue. ` +
    `Return JSON: {"files": string[]}`;

  try {
    const client = getAnthropic();
    const msg = await client.messages.create({
      model: MODEL_SONNET,
      max_tokens: 512,
      temperature: 0,
      messages: [{ role: "user", content: prompt }],
    });
    const text = extractText(msg);
    const parsed = parseJsonResponse<{ files: string[] }>(text);
    if (parsed && Array.isArray(parsed.files)) {
      return parsed.files.slice(0, 20).filter((f: string) => typeof f === "string");
    }
  } catch {
    /* fall through — return empty */
  }
  // Fallback: first 10 files
  return fileTree.slice(0, 10);
}

const MAX_FILE_BYTES = 8 * 1024;       // 8 KB per file
const MAX_TOTAL_BYTES = 80 * 1024;     // 80 KB total

async function readFileContents(
  ownerName: string,
  repoName: string,
  filePaths: string[]
): Promise<Array<{ path: string; content: string }>> {
  const results: Array<{ path: string; content: string }> = [];
  let totalBytes = 0;

  for (const filePath of filePaths) {
    if (totalBytes >= MAX_TOTAL_BYTES) break;
    try {
      const blob = await getBlob(ownerName, repoName, "HEAD", filePath);
      if (!blob || blob.isBinary) continue;
      const content = blob.content.slice(0, MAX_FILE_BYTES);
      totalBytes += content.length;
      results.push({ path: filePath, content });
    } catch {
      /* skip unreadable files */
    }
  }

  return results;
}

interface FilePlan {
  path: string;
  action: "create" | "modify" | "delete";
  description: string;
  patch: string;
}

interface WorkspacePlan {
  summary: string;
  files: FilePlan[];
  branchName: string;
}

async function generatePlan(
  ctx: IssueContext,
  fileContents: Array<{ path: string; content: string }>
): Promise<WorkspacePlan> {
  const fileSection = fileContents
    .map((f) => `=== ${f.path} ===\n${f.content}`)
    .join("\n\n");

  const commentSection = ctx.comments.length > 0
    ? `\n\nComments:\n${ctx.comments.map((c, i) => `[${i + 1}] ${c}`).join("\n\n")}`
    : "";

  const userPrompt =
    `Issue: ${ctx.title}\n${ctx.body}${commentSection}\n\n` +
    `Relevant code:\n${fileSection}\n\n` +
    `Return JSON:\n` +
    `{\n` +
    `  "summary": string,\n` +
    `  "files": Array<{\n` +
    `    "path": string,\n` +
    `    "action": "create" | "modify" | "delete",\n` +
    `    "description": string,\n` +
    `    "patch": string\n` +
    `  }>,\n` +
    `  "branchName": string\n` +
    `}`;

  const systemPrompt =
    "You are an expert software engineer. Generate a detailed implementation plan for the given issue. " +
    "For each file, provide the full new content (not a diff) in the 'patch' field. " +
    "Branch name should follow the pattern: workspace/issue-<number>-<short-slug>. " +
    "Return only valid JSON — no prose, no markdown fences.";

  const client = getAnthropic();
  const msg = await client.messages.create({
    model: MODEL_SONNET,
    max_tokens: 4096,
    temperature: 0.2,
    system: systemPrompt,
    messages: [{ role: "user", content: userPrompt }],
  });

  const text = extractText(msg);
  const parsed = parseJsonResponse<WorkspacePlan>(text);

  if (!parsed || typeof parsed.summary !== "string") {
    throw new Error("AI returned invalid plan JSON");
  }

  return {
    summary: parsed.summary || "AI-generated implementation",
    files: Array.isArray(parsed.files) ? parsed.files : [],
    branchName: typeof parsed.branchName === "string" && parsed.branchName
      ? sanitizeBranchName(parsed.branchName)
      : "",
  };
}

function buildEdits(plan: WorkspacePlan): FileEdit[] {
  const edits: FileEdit[] = [];
  for (const f of plan.files) {
    if (!f.path || typeof f.path !== "string") continue;
    // Reject traversal attempts
    if (f.path.startsWith("/") || f.path.includes("..")) continue;

    if (f.action === "delete") {
      edits.push({ action: "delete", path: f.path });
    } else {
      // create or modify — we always write the full content
      const content = typeof f.patch === "string" ? f.patch : "";
      edits.push({ action: f.action === "create" ? "create" : "edit", path: f.path, content });
    }
  }
  return edits;
}

function formatPlanComment(plan: WorkspacePlan): string {
  const fileLines = plan.files
    .map((f) => `- **${f.action}** \`${f.path}\` — ${f.description}`)
    .join("\n");

  return (
    `<!-- gluecron:workspace:plan -->\n` +
    `## AI Workspace Plan\n\n` +
    `${plan.summary}\n\n` +
    `**Files to change:**\n` +
    (fileLines || "_No files identified_") +
    `\n\n` +
    `_Implementing now... a draft PR will be opened when complete._`
  );
}

async function postIssueComment(
  issueId: string,
  authorId: string,
  body: string
): Promise<string | null> {
  try {
    const [row] = await db
      .insert(issueComments)
      .values({ issueId, authorId, body })
      .returning({ id: issueComments.id });
    return row?.id ?? null;
  } catch {
    return null;
  }
}

async function persistJobStatus(job: WorkspaceJob): Promise<void> {
  try {
    await db
      .update(workspaceJobs)
      .set({
        status: job.status,
        planComment: job.planComment ?? null,
        branchName: job.branchName ?? null,
        prNumber: job.prNumber ?? null,
        errorMessage: job.errorMessage ?? null,
        updatedAt: job.updatedAt,
      })
      .where(eq(workspaceJobs.id, job.id));
  } catch {
    /* non-fatal */
  }
}

// ---------------------------------------------------------------------------
// Utilities
// ---------------------------------------------------------------------------

function slugify(text: string): string {
  return text
    .toLowerCase()
    .replace(/[^a-z0-9]+/g, "-")
    .replace(/^-+|-+$/g, "")
    .slice(0, 40) || "change";
}

function sanitizeBranchName(name: string): string {
  return name
    .replace(/[^a-zA-Z0-9/_-]/g, "-")
    .replace(/--+/g, "-")
    .replace(/^-+|-+$/g, "")
    .slice(0, 100);
}