assistme 0.6.1 → 0.6.3

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
@@ -288,6 +288,7 @@ var SELF_ANALYSIS_MAX_MESSAGE_EVENTS = 300;
288
288
  var SELF_ANALYSIS_MAX_CONVERSATION_MESSAGES = 10;
289
289
  var SELF_ANALYSIS_LOG_CONTEXT_CHARS = 2e4;
290
290
  var SELF_ANALYSIS_EVENT_CONTEXT_CHARS = 2e4;
291
+ var SELF_ANALYSIS_TIMEOUT_MS = 18e4;
291
292
  var EDSGER_PRODUCT_SLUG = "assistme";
292
293
  var MAX_COMPLETE_TASK_RETRIES = 2;
293
294
 
@@ -506,6 +507,7 @@ export {
506
507
  SELF_ANALYSIS_MAX_CONVERSATION_MESSAGES,
507
508
  SELF_ANALYSIS_LOG_CONTEXT_CHARS,
508
509
  SELF_ANALYSIS_EVENT_CONTEXT_CHARS,
510
+ SELF_ANALYSIS_TIMEOUT_MS,
509
511
  EDSGER_PRODUCT_SLUG,
510
512
  MAX_COMPLETE_TASK_RETRIES,
511
513
  AppError,
package/dist/index.js CHANGED
@@ -26,6 +26,7 @@ import {
26
26
  SELF_ANALYSIS_MAX_CONVERSATION_MESSAGES,
27
27
  SELF_ANALYSIS_MAX_MESSAGE_EVENTS,
28
28
  SELF_ANALYSIS_MAX_SESSION_LOGS,
29
+ SELF_ANALYSIS_TIMEOUT_MS,
29
30
  SHELL_MAX_OUTPUT,
30
31
  SHELL_TIMEOUT_MS,
31
32
  SKILL_DESCRIPTION_BUDGET_CHARS,
@@ -45,7 +46,7 @@ import {
45
46
  setLogHook,
46
47
  setLogLevel,
47
48
  writeAuthStore
48
- } from "./chunk-ECEOBNDM.js";
49
+ } from "./chunk-3V6TCGZG.js";
49
50
  import {
50
51
  clearConfig,
51
52
  getConfig,
@@ -3845,11 +3846,11 @@ var SELF_ANALYSIS_OUTPUT_FORMAT = {
3845
3846
  type: "object",
3846
3847
  properties: {
3847
3848
  is_perfect: { type: "boolean" },
3848
- overall_score: { type: "number" },
3849
+ overall_score: { type: "number", minimum: 1, maximum: 10 },
3849
3850
  task_completion_quality: {
3850
3851
  type: "object",
3851
3852
  properties: {
3852
- score: { type: "number" },
3853
+ score: { type: "number", minimum: 1, maximum: 10 },
3853
3854
  assessment: { type: "string" }
3854
3855
  },
3855
3856
  required: ["score", "assessment"]
@@ -3874,11 +3875,11 @@ var SELF_ANALYSIS_OUTPUT_FORMAT = {
3874
3875
  type: "object",
3875
3876
  properties: {
3876
3877
  session_logs_useful: { type: "boolean" },
3877
- session_logs_gaps: { type: "string" },
3878
+ session_logs_gaps: { type: ["string", "null"] },
3878
3879
  message_events_useful: { type: "boolean" },
3879
- message_events_gaps: { type: "string" },
3880
+ message_events_gaps: { type: ["string", "null"] },
3880
3881
  conversation_context_useful: { type: "boolean" },
3881
- conversation_context_gaps: { type: "string" }
3882
+ conversation_context_gaps: { type: ["string", "null"] }
3882
3883
  },
3883
3884
  required: [
3884
3885
  "session_logs_useful",
@@ -4104,6 +4105,31 @@ ${dataQualityNotes}
4104
4105
  }
4105
4106
  }
4106
4107
  }
4108
+ async function runAnalysisQuery(model, prompt) {
4109
+ let structuredOutput;
4110
+ for await (const message of query2({
4111
+ prompt,
4112
+ options: {
4113
+ model,
4114
+ maxTurns: 1,
4115
+ allowedTools: [],
4116
+ effort: "low",
4117
+ outputFormat: SELF_ANALYSIS_OUTPUT_FORMAT
4118
+ }
4119
+ })) {
4120
+ if (message.type === "result") {
4121
+ const resultMsg = message;
4122
+ if (resultMsg.subtype === "success") {
4123
+ const successMsg = resultMsg;
4124
+ structuredOutput = successMsg.structured_output;
4125
+ log.debug(
4126
+ `Self-analysis cost: $${successMsg.total_cost_usd.toFixed(4)}`
4127
+ );
4128
+ }
4129
+ }
4130
+ }
4131
+ return structuredOutput;
4132
+ }
4107
4133
  async function analyzeSelfPostTask(opts) {
4108
4134
  const {
4109
4135
  model,
@@ -4132,31 +4158,25 @@ async function analyzeSelfPostTask(opts) {
4132
4158
  ${analysisContext}
4133
4159
 
4134
4160
  Respond with a JSON object now.`;
4135
- let structuredOutput;
4136
- for await (const message of query2({
4137
- prompt,
4138
- options: {
4139
- model,
4140
- maxTurns: 1,
4141
- allowedTools: [],
4142
- effort: "low",
4143
- outputFormat: SELF_ANALYSIS_OUTPUT_FORMAT
4144
- }
4145
- })) {
4146
- if (message.type === "result") {
4147
- const resultMsg = message;
4148
- if (resultMsg.subtype === "success") {
4149
- const successMsg = resultMsg;
4150
- structuredOutput = successMsg.structured_output;
4151
- log.debug(
4152
- `Self-analysis cost: $${successMsg.total_cost_usd.toFixed(4)}`
4153
- );
4154
- }
4161
+ const analysisPromise = runAnalysisQuery(model, prompt);
4162
+ const timeoutPromise = new Promise(
4163
+ (_, reject) => setTimeout(() => reject(new Error(`Self-analysis timed out after ${SELF_ANALYSIS_TIMEOUT_MS / 1e3}s`)), SELF_ANALYSIS_TIMEOUT_MS)
4164
+ );
4165
+ const structuredOutput = await Promise.race([analysisPromise, timeoutPromise]);
4166
+ let analysis = null;
4167
+ if (structuredOutput) {
4168
+ const result = SelfAnalysisResultSchema.safeParse(structuredOutput);
4169
+ if (result.success) {
4170
+ analysis = result.data;
4171
+ } else {
4172
+ log.warn(
4173
+ `Self-analysis: schema validation failed: ${result.error.issues.map((i) => `${i.path.join(".")}: ${i.message}`).join("; ")}`
4174
+ );
4175
+ log.debug(`Self-analysis: raw output: ${JSON.stringify(structuredOutput).slice(0, 500)}`);
4155
4176
  }
4156
4177
  }
4157
- const analysis = structuredOutput ? safeParse(SelfAnalysisResultSchema, structuredOutput) : null;
4158
4178
  if (!analysis) {
4159
- log.debug("Self-analysis: no valid structured output");
4179
+ log.warn("Self-analysis: no valid structured output");
4160
4180
  return;
4161
4181
  }
4162
4182
  log.info(
@@ -4168,7 +4188,7 @@ Respond with a JSON object now.`;
4168
4188
  log.debug("Self-analysis: no improvements to report \u2014 skipping feedback");
4169
4189
  }
4170
4190
  } catch (err) {
4171
- log.debug(`Self-analysis error: ${errorMessage(err)}`);
4191
+ log.warn(`Self-analysis error: ${errorMessage(err)}`);
4172
4192
  }
4173
4193
  }
4174
4194
 
@@ -7075,7 +7095,7 @@ function registerJobCommands(program2) {
7075
7095
  jobCmd.command("list").description("List your defined jobs").action(async () => {
7076
7096
  try {
7077
7097
  const userId = await getCurrentUserId();
7078
- const { JobRunner: JobRunner2 } = await import("./job-runner-RGP4CLYV.js");
7098
+ const { JobRunner: JobRunner2 } = await import("./job-runner-W6CTTJMR.js");
7079
7099
  const runner = new JobRunner2();
7080
7100
  const jobs = await runner.listJobs();
7081
7101
  if (jobs.length === 0) {
@@ -7099,7 +7119,7 @@ function registerJobCommands(program2) {
7099
7119
  jobCmd.command("status [name]").description("Show run history for a job (or all jobs)").option("-l, --limit <number>", "Max runs to show (default: 5)").action(async (name, opts) => {
7100
7120
  try {
7101
7121
  const userId = await getCurrentUserId();
7102
- const { JobRunner: JobRunner2 } = await import("./job-runner-RGP4CLYV.js");
7122
+ const { JobRunner: JobRunner2 } = await import("./job-runner-W6CTTJMR.js");
7103
7123
  const runner = new JobRunner2();
7104
7124
  const runs = await runner.getRunHistory(name, parseInt(opts.limit || "5"));
7105
7125
  if (runs.length === 0) {
@@ -7138,7 +7158,7 @@ Job Run History${name ? ` \u2014 ${name}` : ""}:`));
7138
7158
  process.exit(1);
7139
7159
  }
7140
7160
  const userId = await getCurrentUserId();
7141
- const { JobRunner: JobRunner2 } = await import("./job-runner-RGP4CLYV.js");
7161
+ const { JobRunner: JobRunner2 } = await import("./job-runner-W6CTTJMR.js");
7142
7162
  const runner = new JobRunner2();
7143
7163
  const job = await runner.loadJob(name);
7144
7164
  if (!job) {
@@ -1,6 +1,6 @@
1
1
  import {
2
2
  JobRunner
3
- } from "./chunk-ECEOBNDM.js";
3
+ } from "./chunk-3V6TCGZG.js";
4
4
  import "./chunk-JVA6DHXD.js";
5
5
  export {
6
6
  JobRunner
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "assistme",
3
- "version": "0.6.1",
3
+ "version": "0.6.3",
4
4
  "description": "AssistMe CLI Agent - AI-powered assistant that controls your real browser",
5
5
  "type": "module",
6
6
  "main": "dist/index.js",
@@ -9,7 +9,6 @@ import { log } from "../utils/logger.js";
9
9
  import {
10
10
  SelfAnalysisResultSchema,
11
11
  type SelfAnalysisResult,
12
- safeParse,
13
12
  } from "../utils/schemas.js";
14
13
  import { errorMessage } from "../utils/errors.js";
15
14
  import {
@@ -23,6 +22,7 @@ import {
23
22
  SELF_ANALYSIS_MAX_CONVERSATION_MESSAGES,
24
23
  SELF_ANALYSIS_LOG_CONTEXT_CHARS,
25
24
  SELF_ANALYSIS_EVENT_CONTEXT_CHARS,
25
+ SELF_ANALYSIS_TIMEOUT_MS,
26
26
  EDSGER_PRODUCT_SLUG,
27
27
  } from "../utils/constants.js";
28
28
  import type { ToolCallRecord } from "./skill-extractor.js";
@@ -36,11 +36,11 @@ const SELF_ANALYSIS_OUTPUT_FORMAT: OutputFormat = {
36
36
  type: "object",
37
37
  properties: {
38
38
  is_perfect: { type: "boolean" },
39
- overall_score: { type: "number" },
39
+ overall_score: { type: "number", minimum: 1, maximum: 10 },
40
40
  task_completion_quality: {
41
41
  type: "object",
42
42
  properties: {
43
- score: { type: "number" },
43
+ score: { type: "number", minimum: 1, maximum: 10 },
44
44
  assessment: { type: "string" },
45
45
  },
46
46
  required: ["score", "assessment"],
@@ -65,11 +65,11 @@ const SELF_ANALYSIS_OUTPUT_FORMAT: OutputFormat = {
65
65
  type: "object",
66
66
  properties: {
67
67
  session_logs_useful: { type: "boolean" },
68
- session_logs_gaps: { type: "string" },
68
+ session_logs_gaps: { type: ["string", "null"] },
69
69
  message_events_useful: { type: "boolean" },
70
- message_events_gaps: { type: "string" },
70
+ message_events_gaps: { type: ["string", "null"] },
71
71
  conversation_context_useful: { type: "boolean" },
72
- conversation_context_gaps: { type: "string" },
72
+ conversation_context_gaps: { type: ["string", "null"] },
73
73
  },
74
74
  required: [
75
75
  "session_logs_useful",
@@ -345,6 +345,38 @@ async function submitSelfAnalysisFeedback(analysis: SelfAnalysisResult): Promise
345
345
  }
346
346
  }
347
347
 
348
+ // ── Query Runner ────────────────────────────────────────────────
349
+
350
+ async function runAnalysisQuery(model: string, prompt: string): Promise<unknown> {
351
+ let structuredOutput: unknown;
352
+
353
+ // Use independent query() instead of session resume to avoid
354
+ // conflicts with skill evaluation which also resumes the session
355
+ for await (const message of query({
356
+ prompt,
357
+ options: {
358
+ model,
359
+ maxTurns: 1,
360
+ allowedTools: [],
361
+ effort: "low",
362
+ outputFormat: SELF_ANALYSIS_OUTPUT_FORMAT,
363
+ },
364
+ })) {
365
+ if (message.type === "result") {
366
+ const resultMsg = message as SDKResultMessage;
367
+ if (resultMsg.subtype === "success") {
368
+ const successMsg = resultMsg as SDKResultSuccess;
369
+ structuredOutput = successMsg.structured_output;
370
+ log.debug(
371
+ `Self-analysis cost: $${successMsg.total_cost_usd.toFixed(4)}`
372
+ );
373
+ }
374
+ }
375
+ }
376
+
377
+ return structuredOutput;
378
+ }
379
+
348
380
  // ── Main Entry Point ────────────────────────────────────────────
349
381
 
350
382
  /**
@@ -394,39 +426,30 @@ export async function analyzeSelfPostTask(opts: {
394
426
 
395
427
  const prompt = `${SELF_ANALYSIS_PROMPT}\n${analysisContext}\n\nRespond with a JSON object now.`;
396
428
 
397
- let structuredOutput: unknown;
398
-
399
- // Use independent query() instead of session resume to avoid
400
- // conflicts with skill evaluation which also resumes the session
401
- for await (const message of query({
402
- prompt,
403
- options: {
404
- model,
405
- maxTurns: 1,
406
- allowedTools: [],
407
- effort: "low",
408
- outputFormat: SELF_ANALYSIS_OUTPUT_FORMAT,
409
- },
410
- })) {
411
- if (message.type === "result") {
412
- const resultMsg = message as SDKResultMessage;
413
- if (resultMsg.subtype === "success") {
414
- const successMsg = resultMsg as SDKResultSuccess;
415
- structuredOutput = successMsg.structured_output;
416
- log.debug(
417
- `Self-analysis cost: $${successMsg.total_cost_usd.toFixed(4)}`
418
- );
419
- }
420
- }
421
- }
429
+ // Race the analysis against a timeout to avoid hanging forever
430
+ const analysisPromise = runAnalysisQuery(model, prompt);
431
+ const timeoutPromise = new Promise<never>((_, reject) =>
432
+ setTimeout(() => reject(new Error(`Self-analysis timed out after ${SELF_ANALYSIS_TIMEOUT_MS / 1000}s`)), SELF_ANALYSIS_TIMEOUT_MS)
433
+ );
434
+
435
+ const structuredOutput = await Promise.race([analysisPromise, timeoutPromise]);
422
436
 
423
437
  // Validate against Zod schema
424
- const analysis = structuredOutput
425
- ? safeParse(SelfAnalysisResultSchema, structuredOutput)
426
- : null;
438
+ let analysis: SelfAnalysisResult | null = null;
439
+ if (structuredOutput) {
440
+ const result = SelfAnalysisResultSchema.safeParse(structuredOutput);
441
+ if (result.success) {
442
+ analysis = result.data;
443
+ } else {
444
+ log.warn(
445
+ `Self-analysis: schema validation failed: ${result.error.issues.map((i) => `${i.path.join(".")}: ${i.message}`).join("; ")}`
446
+ );
447
+ log.debug(`Self-analysis: raw output: ${JSON.stringify(structuredOutput).slice(0, 500)}`);
448
+ }
449
+ }
427
450
 
428
451
  if (!analysis) {
429
- log.debug("Self-analysis: no valid structured output");
452
+ log.warn("Self-analysis: no valid structured output");
430
453
  return;
431
454
  }
432
455
 
@@ -441,6 +464,6 @@ export async function analyzeSelfPostTask(opts: {
441
464
  log.debug("Self-analysis: no improvements to report — skipping feedback");
442
465
  }
443
466
  } catch (err) {
444
- log.debug(`Self-analysis error: ${errorMessage(err)}`);
467
+ log.warn(`Self-analysis error: ${errorMessage(err)}`);
445
468
  }
446
469
  }
@@ -106,6 +106,9 @@ export const SELF_ANALYSIS_LOG_CONTEXT_CHARS = 20_000;
106
106
  /** Max characters for event context in self-analysis */
107
107
  export const SELF_ANALYSIS_EVENT_CONTEXT_CHARS = 20_000;
108
108
 
109
+ /** Timeout for the self-analysis query in ms (3 minutes) */
110
+ export const SELF_ANALYSIS_TIMEOUT_MS = 180_000;
111
+
109
112
  /** Edsger feedback product slug for assistme */
110
113
  export const EDSGER_PRODUCT_SLUG = "assistme";
111
114