opc-agent 1.4.0 → 1.4.1

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.
Files changed (192) hide show
  1. package/CHANGELOG.md +69 -23
  2. package/CONTRIBUTING.md +60 -21
  3. package/README.md +358 -235
  4. package/README.zh-CN.md +415 -415
  5. package/dist/channels/slack.js +10 -93
  6. package/dist/channels/web.d.ts +0 -10
  7. package/dist/channels/web.js +2 -33
  8. package/dist/cli.js +60 -255
  9. package/dist/core/dashboard.d.ts +35 -0
  10. package/dist/core/dashboard.js +157 -0
  11. package/dist/core/fast-mode.d.ts +27 -0
  12. package/dist/core/fast-mode.js +59 -0
  13. package/dist/core/priority.d.ts +52 -0
  14. package/dist/core/priority.js +102 -0
  15. package/dist/core/runtime.d.ts +0 -4
  16. package/dist/core/runtime.js +0 -27
  17. package/dist/deploy/hermes.js +22 -22
  18. package/dist/deploy/openclaw.js +40 -31
  19. package/dist/index.d.ts +14 -3
  20. package/dist/index.js +20 -6
  21. package/dist/memory/cloud-storage.d.ts +40 -0
  22. package/dist/memory/cloud-storage.js +211 -0
  23. package/dist/providers/index.d.ts +1 -1
  24. package/dist/providers/index.js +1 -7
  25. package/dist/schema/oad.d.ts +2 -1
  26. package/dist/templates/code-reviewer.d.ts +8 -0
  27. package/dist/templates/code-reviewer.js +9 -5
  28. package/dist/templates/customer-service.d.ts +8 -0
  29. package/dist/templates/customer-service.js +6 -2
  30. package/dist/templates/data-analyst.d.ts +8 -0
  31. package/dist/templates/data-analyst.js +9 -5
  32. package/dist/templates/knowledge-base.d.ts +8 -0
  33. package/dist/templates/knowledge-base.js +6 -2
  34. package/dist/templates/sales-assistant.d.ts +8 -0
  35. package/dist/templates/sales-assistant.js +8 -4
  36. package/dist/templates/teacher.d.ts +8 -0
  37. package/dist/templates/teacher.js +10 -6
  38. package/docs/.vitepress/config.ts +103 -103
  39. package/docs/api/cli.md +48 -48
  40. package/docs/api/oad-schema.md +64 -64
  41. package/docs/api/sdk.md +80 -80
  42. package/docs/guide/concepts.md +51 -51
  43. package/docs/guide/configuration.md +79 -79
  44. package/docs/guide/deployment.md +42 -42
  45. package/docs/guide/getting-started.md +44 -44
  46. package/docs/guide/templates.md +28 -28
  47. package/docs/guide/testing.md +84 -84
  48. package/docs/index.md +27 -27
  49. package/docs/zh/api/cli.md +54 -54
  50. package/docs/zh/api/oad-schema.md +87 -87
  51. package/docs/zh/api/sdk.md +102 -102
  52. package/docs/zh/guide/concepts.md +104 -104
  53. package/docs/zh/guide/configuration.md +135 -135
  54. package/docs/zh/guide/deployment.md +81 -81
  55. package/docs/zh/guide/getting-started.md +82 -82
  56. package/docs/zh/guide/templates.md +84 -84
  57. package/docs/zh/guide/testing.md +88 -88
  58. package/docs/zh/index.md +27 -27
  59. package/examples/customer-service-demo/README.md +90 -90
  60. package/examples/customer-service-demo/oad.yaml +107 -107
  61. package/package.json +1 -1
  62. package/src/analytics/index.ts +66 -66
  63. package/src/channels/discord.ts +192 -192
  64. package/src/channels/email.ts +177 -177
  65. package/src/channels/feishu.ts +236 -236
  66. package/src/channels/index.ts +15 -15
  67. package/src/channels/slack.ts +160 -217
  68. package/src/channels/telegram.ts +90 -90
  69. package/src/channels/voice.ts +106 -106
  70. package/src/channels/web.ts +2 -38
  71. package/src/channels/webhook.ts +199 -199
  72. package/src/channels/websocket.ts +87 -87
  73. package/src/channels/wechat.ts +149 -149
  74. package/src/cli.ts +58 -282
  75. package/src/core/a2a.ts +143 -143
  76. package/src/core/agent.ts +152 -152
  77. package/src/core/analytics-engine.ts +186 -186
  78. package/src/core/auth.ts +57 -57
  79. package/src/core/cache.ts +141 -141
  80. package/src/core/compose.ts +77 -77
  81. package/src/core/config.ts +14 -14
  82. package/src/core/dashboard.ts +219 -0
  83. package/src/core/errors.ts +148 -148
  84. package/src/core/fast-mode.ts +75 -0
  85. package/src/core/hitl.ts +138 -138
  86. package/src/core/logger.ts +57 -57
  87. package/src/core/orchestrator.ts +215 -215
  88. package/src/core/performance.ts +187 -187
  89. package/src/core/priority.ts +140 -0
  90. package/src/core/rate-limiter.ts +128 -128
  91. package/src/core/room.ts +109 -109
  92. package/src/core/runtime.ts +152 -183
  93. package/src/core/sandbox.ts +101 -101
  94. package/src/core/security.ts +171 -171
  95. package/src/core/types.ts +68 -68
  96. package/src/core/versioning.ts +106 -106
  97. package/src/core/watch.ts +178 -178
  98. package/src/core/workflow.ts +235 -235
  99. package/src/deploy/hermes.ts +156 -156
  100. package/src/deploy/openclaw.ts +200 -190
  101. package/src/dtv/data.ts +29 -0
  102. package/src/dtv/trust.ts +43 -0
  103. package/src/dtv/value.ts +47 -0
  104. package/src/i18n/index.ts +216 -216
  105. package/src/index.ts +16 -3
  106. package/src/marketplace/index.ts +223 -0
  107. package/src/memory/cloud-storage.ts +217 -0
  108. package/src/memory/deepbrain.ts +108 -108
  109. package/src/memory/index.ts +34 -34
  110. package/src/plugins/index.ts +208 -208
  111. package/src/providers/index.ts +1 -9
  112. package/src/schema/oad.ts +155 -154
  113. package/src/skills/base.ts +16 -16
  114. package/src/skills/document.ts +100 -100
  115. package/src/skills/http.ts +35 -35
  116. package/src/skills/index.ts +27 -27
  117. package/src/skills/scheduler.ts +80 -80
  118. package/src/skills/webhook-trigger.ts +59 -59
  119. package/src/templates/code-reviewer.ts +34 -30
  120. package/src/templates/customer-service.ts +80 -76
  121. package/src/templates/data-analyst.ts +70 -66
  122. package/src/templates/executive-assistant.ts +71 -71
  123. package/src/templates/financial-advisor.ts +60 -60
  124. package/src/templates/knowledge-base.ts +31 -27
  125. package/src/templates/legal-assistant.ts +71 -71
  126. package/src/templates/sales-assistant.ts +79 -75
  127. package/src/templates/teacher.ts +79 -75
  128. package/src/testing/index.ts +181 -181
  129. package/src/tools/calculator.ts +73 -73
  130. package/src/tools/datetime.ts +149 -149
  131. package/src/tools/json-transform.ts +187 -187
  132. package/src/tools/mcp.ts +76 -76
  133. package/src/tools/text-analysis.ts +116 -116
  134. package/templates/Dockerfile +15 -15
  135. package/templates/code-reviewer/README.md +27 -27
  136. package/templates/code-reviewer/oad.yaml +41 -41
  137. package/templates/customer-service/README.md +22 -22
  138. package/templates/customer-service/oad.yaml +36 -36
  139. package/templates/docker-compose.yml +21 -21
  140. package/templates/ecommerce-assistant/README.md +45 -45
  141. package/templates/ecommerce-assistant/oad.yaml +47 -47
  142. package/templates/knowledge-base/README.md +28 -28
  143. package/templates/knowledge-base/oad.yaml +38 -38
  144. package/templates/sales-assistant/README.md +26 -26
  145. package/templates/sales-assistant/oad.yaml +43 -43
  146. package/templates/tech-support/README.md +43 -43
  147. package/templates/tech-support/oad.yaml +45 -45
  148. package/tests/a2a.test.ts +66 -66
  149. package/tests/agent.test.ts +72 -72
  150. package/tests/analytics.test.ts +50 -50
  151. package/tests/channel.test.ts +39 -39
  152. package/tests/e2e.test.ts +134 -134
  153. package/tests/errors.test.ts +83 -83
  154. package/tests/hitl.test.ts +71 -71
  155. package/tests/i18n.test.ts +41 -41
  156. package/tests/mcp.test.ts +54 -54
  157. package/tests/oad.test.ts +68 -68
  158. package/tests/performance.test.ts +115 -115
  159. package/tests/plugin.test.ts +74 -74
  160. package/tests/room.test.ts +106 -106
  161. package/tests/runtime.test.ts +42 -42
  162. package/tests/sandbox.test.ts +46 -46
  163. package/tests/security.test.ts +60 -60
  164. package/tests/templates.test.ts +77 -77
  165. package/tests/v070.test.ts +76 -76
  166. package/tests/versioning.test.ts +75 -75
  167. package/tests/voice.test.ts +61 -61
  168. package/tests/webhook.test.ts +29 -29
  169. package/tests/workflow.test.ts +143 -143
  170. package/tsconfig.json +19 -19
  171. package/vitest.config.ts +9 -9
  172. package/.github/ISSUE_TEMPLATE/bug_report.md +0 -20
  173. package/.github/ISSUE_TEMPLATE/feature_request.md +0 -14
  174. package/.github/PULL_REQUEST_TEMPLATE.md +0 -13
  175. package/.github/workflows/ci.yml +0 -24
  176. package/dist/traces/index.d.ts +0 -49
  177. package/dist/traces/index.js +0 -102
  178. package/examples/README.md +0 -22
  179. package/examples/basic-agent.ts +0 -90
  180. package/examples/brain-integration.ts +0 -71
  181. package/examples/multi-channel.ts +0 -74
  182. package/src/traces/index.ts +0 -132
  183. package/test-agent/Dockerfile +0 -9
  184. package/test-agent/README.md +0 -50
  185. package/test-agent/agent.yaml +0 -23
  186. package/test-agent/docker-compose.yml +0 -11
  187. package/test-agent/oad.yaml +0 -31
  188. package/test-agent/package-lock.json +0 -1492
  189. package/test-agent/package.json +0 -18
  190. package/test-agent/src/index.ts +0 -24
  191. package/test-agent/src/skills/echo.ts +0 -15
  192. package/test-agent/tsconfig.json +0 -25
@@ -1,75 +1,79 @@
1
- import { BaseSkill } from '../skills/base';
2
- import type { AgentContext, Message, SkillResult } from '../core/types';
3
-
4
- export class LessonPlanSkill extends BaseSkill {
5
- name = 'lesson-plan';
6
- description = 'Create and manage lesson plans';
7
-
8
- async execute(_context: AgentContext, message: Message): Promise<SkillResult> {
9
- const lower = message.content.toLowerCase();
10
- if (lower.includes('lesson') || lower.includes('plan') || lower.includes('curriculum') || lower.includes('syllabus')) {
11
- return this.match('I can help create a lesson plan. What subject, grade level, and learning objectives should I focus on?', 0.85);
12
- }
13
- return this.noMatch();
14
- }
15
- }
16
-
17
- export class QuizSkill extends BaseSkill {
18
- name = 'quiz-generator';
19
- description = 'Generate quizzes and assessments';
20
-
21
- async execute(_context: AgentContext, message: Message): Promise<SkillResult> {
22
- const lower = message.content.toLowerCase();
23
- if (lower.includes('quiz') || lower.includes('test') || lower.includes('assessment') || lower.includes('exam') || lower.includes('question')) {
24
- return this.match('I\'ll create a quiz for you. What topic, difficulty level, and number of questions would you like?', 0.85);
25
- }
26
- return this.noMatch();
27
- }
28
- }
29
-
30
- export class ExplainSkill extends BaseSkill {
31
- name = 'concept-explainer';
32
- description = 'Explain concepts at appropriate level';
33
-
34
- async execute(_context: AgentContext, message: Message): Promise<SkillResult> {
35
- const lower = message.content.toLowerCase();
36
- if (lower.includes('explain') || lower.includes('what is') || lower.includes('how does') || lower.includes('why')) {
37
- return this.match('Let me explain that concept. What\'s your current level of understanding so I can tailor my explanation?', 0.75);
38
- }
39
- return this.noMatch();
40
- }
41
- }
42
-
43
- export const TEACHER_SYSTEM_PROMPT = `You are a patient and encouraging teacher assistant. Your goals:
44
- 1. Create engaging lesson plans tailored to student level
45
- 2. Generate quizzes and assessments with answer keys
46
- 3. Explain complex concepts using analogies and examples
47
- 4. Provide constructive feedback and encouragement
48
- 5. Adapt teaching style to different learning preferences
49
- Be patient, use clear language, and always check for understanding. Use the Socratic method when appropriate.`;
50
-
51
- export function createTeacherConfig() {
52
- return {
53
- apiVersion: 'opc/v1' as const,
54
- kind: 'Agent' as const,
55
- metadata: {
56
- name: 'teacher',
57
- version: '1.0.0',
58
- description: 'AI teacher assistant with lesson planning, quiz generation, and concept explanation',
59
- author: 'OPC Agent',
60
- license: 'Apache-2.0',
61
- },
62
- spec: {
63
- provider: { default: 'openai', allowed: ['openai', 'deepseek', 'qwen'] },
64
- model: 'gpt-4o-mini',
65
- systemPrompt: TEACHER_SYSTEM_PROMPT,
66
- skills: [
67
- { name: 'lesson-plan', description: 'Create lesson plans' },
68
- { name: 'quiz-generator', description: 'Generate quizzes' },
69
- { name: 'concept-explainer', description: 'Explain concepts' },
70
- ],
71
- channels: [{ type: 'web' as const, port: 3000 }],
72
- memory: { shortTerm: true, longTerm: true },
73
- },
74
- };
75
- }
1
+ import { BaseSkill } from '../skills/base';
2
+ import type { AgentContext, Message, SkillResult } from '../core/types';
3
+
4
+ export class LessonPlanSkill extends BaseSkill {
5
+ name = 'lesson-plan';
6
+ description = 'Create and manage lesson plans';
7
+
8
+ async execute(_context: AgentContext, message: Message): Promise<SkillResult> {
9
+ const lower = message.content.toLowerCase();
10
+ if (lower.includes('lesson') || lower.includes('plan') || lower.includes('curriculum') || lower.includes('syllabus')) {
11
+ return this.match('I can help create a lesson plan. What subject, grade level, and learning objectives should I focus on?', 0.85);
12
+ }
13
+ return this.noMatch();
14
+ }
15
+ }
16
+
17
+ export class QuizSkill extends BaseSkill {
18
+ name = 'quiz-generator';
19
+ description = 'Generate quizzes and assessments';
20
+
21
+ async execute(_context: AgentContext, message: Message): Promise<SkillResult> {
22
+ const lower = message.content.toLowerCase();
23
+ if (lower.includes('quiz') || lower.includes('test') || lower.includes('assessment') || lower.includes('exam') || lower.includes('question')) {
24
+ return this.match('I\'ll create a quiz for you. What topic, difficulty level, and number of questions would you like?', 0.85);
25
+ }
26
+ return this.noMatch();
27
+ }
28
+ }
29
+
30
+ export class ExplainSkill extends BaseSkill {
31
+ name = 'concept-explainer';
32
+ description = 'Explain concepts at appropriate level';
33
+
34
+ async execute(_context: AgentContext, message: Message): Promise<SkillResult> {
35
+ const lower = message.content.toLowerCase();
36
+ if (lower.includes('explain') || lower.includes('what is') || lower.includes('how does') || lower.includes('why')) {
37
+ return this.match('Let me explain that concept. What\'s your current level of understanding so I can tailor my explanation?', 0.75);
38
+ }
39
+ return this.noMatch();
40
+ }
41
+ }
42
+
43
+ export const TEACHER_SYSTEM_PROMPT = `You are a patient and encouraging teacher assistant. Your goals:
44
+ 1. Create engaging lesson plans tailored to student level
45
+ 2. Generate quizzes and assessments with answer keys
46
+ 3. Explain complex concepts using analogies and examples
47
+ 4. Provide constructive feedback and encouragement
48
+ 5. Adapt teaching style to different learning preferences
49
+ Be patient, use clear language, and always check for understanding. Use the Socratic method when appropriate.`;
50
+
51
+ export function createTeacherConfig() {
52
+ return {
53
+ apiVersion: 'opc/v1' as const,
54
+ kind: 'Agent' as const,
55
+ metadata: {
56
+ name: 'teacher',
57
+ version: '1.0.0',
58
+ description: 'AI teacher assistant with lesson planning, quiz generation, and concept explanation',
59
+ author: 'OPC Agent',
60
+ license: 'Apache-2.0',
61
+ },
62
+ spec: {
63
+ provider: { default: 'openai', allowed: ['openai', 'deepseek', 'qwen'] },
64
+ model: 'gpt-4o-mini',
65
+ systemPrompt: TEACHER_SYSTEM_PROMPT,
66
+ skills: [
67
+ { name: 'lesson-plan', description: 'Create lesson plans' },
68
+ { name: 'quiz-generator', description: 'Generate quizzes' },
69
+ { name: 'concept-explainer', description: 'Explain concepts' },
70
+ ],
71
+ channels: [{ type: 'web' as const, port: 3000 }],
72
+ memory: { shortTerm: true, longTerm: true },
73
+ dtv: {
74
+ trust: { level: 'sandbox' as const },
75
+ value: { metrics: ['lessons_created', 'quizzes_generated', 'concepts_explained'] },
76
+ },
77
+ },
78
+ };
79
+ }
@@ -1,181 +1,181 @@
1
- /**
2
- * Agent Testing Framework - Define test cases in OAD, run with `opc test`.
3
- * Supports assertions on response content, tool calls, and latency.
4
- */
5
- import * as fs from 'fs';
6
- import * as path from 'path';
7
- import * as yaml from 'js-yaml';
8
- import { AgentRuntime } from '../core/runtime';
9
-
10
- export interface TestCase {
11
- name: string;
12
- input: string;
13
- expect?: {
14
- contains?: string[];
15
- notContains?: string[];
16
- toolCalled?: string[];
17
- maxLatencyMs?: number;
18
- };
19
- }
20
-
21
- export interface TestResult {
22
- name: string;
23
- passed: boolean;
24
- durationMs: number;
25
- response?: string;
26
- failures: string[];
27
- }
28
-
29
- export interface TestReport {
30
- total: number;
31
- passed: number;
32
- failed: number;
33
- duration: number;
34
- results: TestResult[];
35
- }
36
-
37
- /**
38
- * Load test cases from OAD spec.testing or a separate test file.
39
- */
40
- export function loadTestCases(oadPath: string): TestCase[] {
41
- const raw = fs.readFileSync(oadPath, 'utf-8');
42
- const config = yaml.load(raw) as any;
43
-
44
- // Check spec.testing.cases
45
- if (config?.spec?.testing?.cases) {
46
- return config.spec.testing.cases;
47
- }
48
-
49
- // Check for companion test file
50
- const dir = path.dirname(oadPath);
51
- const testFile = path.join(dir, 'tests.yaml');
52
- if (fs.existsSync(testFile)) {
53
- const testRaw = fs.readFileSync(testFile, 'utf-8');
54
- const testConfig = yaml.load(testRaw) as any;
55
- return testConfig?.cases ?? testConfig ?? [];
56
- }
57
-
58
- return [];
59
- }
60
-
61
- /**
62
- * Run all test cases against an agent.
63
- */
64
- export async function runTests(oadPath: string): Promise<TestReport> {
65
- const cases = loadTestCases(oadPath);
66
- const results: TestResult[] = [];
67
- const startTime = Date.now();
68
-
69
- if (cases.length === 0) {
70
- // Generate default smoke test
71
- cases.push({
72
- name: 'smoke-test',
73
- input: 'Hello! What can you help me with?',
74
- expect: { maxLatencyMs: 30000 },
75
- });
76
- }
77
-
78
- const runtime = new AgentRuntime();
79
- await runtime.loadConfig(oadPath);
80
- const agent = await runtime.initialize();
81
-
82
- for (const tc of cases) {
83
- const result: TestResult = {
84
- name: tc.name,
85
- passed: true,
86
- durationMs: 0,
87
- failures: [],
88
- };
89
-
90
- const t0 = Date.now();
91
- try {
92
- const response = await agent.handleMessage({
93
- id: `test_${Date.now()}`,
94
- role: 'user',
95
- content: tc.input,
96
- timestamp: Date.now(),
97
- });
98
- result.durationMs = Date.now() - t0;
99
- result.response = response.content;
100
-
101
- if (tc.expect) {
102
- // Check contains
103
- if (tc.expect.contains) {
104
- for (const s of tc.expect.contains) {
105
- if (!response.content.toLowerCase().includes(s.toLowerCase())) {
106
- result.failures.push(`Expected response to contain "${s}"`);
107
- }
108
- }
109
- }
110
- // Check notContains
111
- if (tc.expect.notContains) {
112
- for (const s of tc.expect.notContains) {
113
- if (response.content.toLowerCase().includes(s.toLowerCase())) {
114
- result.failures.push(`Expected response NOT to contain "${s}"`);
115
- }
116
- }
117
- }
118
- // Check latency
119
- if (tc.expect.maxLatencyMs && result.durationMs > tc.expect.maxLatencyMs) {
120
- result.failures.push(`Latency ${result.durationMs}ms exceeded max ${tc.expect.maxLatencyMs}ms`);
121
- }
122
- // Check tool calls (from metadata if available)
123
- if (tc.expect.toolCalled && (response as any).toolsCalled) {
124
- for (const tool of tc.expect.toolCalled) {
125
- if (!(response as any).toolsCalled.includes(tool)) {
126
- result.failures.push(`Expected tool "${tool}" to be called`);
127
- }
128
- }
129
- }
130
- }
131
-
132
- result.passed = result.failures.length === 0;
133
- } catch (err) {
134
- result.durationMs = Date.now() - t0;
135
- result.passed = false;
136
- result.failures.push(`Error: ${err instanceof Error ? err.message : String(err)}`);
137
- }
138
-
139
- results.push(result);
140
- }
141
-
142
- const totalDuration = Date.now() - startTime;
143
- const passed = results.filter(r => r.passed).length;
144
-
145
- return {
146
- total: results.length,
147
- passed,
148
- failed: results.length - passed,
149
- duration: totalDuration,
150
- results,
151
- };
152
- }
153
-
154
- /**
155
- * Format test report for console output.
156
- */
157
- export function formatReport(report: TestReport): string {
158
- const lines: string[] = [];
159
- lines.push('');
160
- lines.push('═══════════════════════════════════════════');
161
- lines.push(' OPC Agent Test Report');
162
- lines.push('═══════════════════════════════════════════');
163
- lines.push('');
164
-
165
- for (const r of report.results) {
166
- const icon = r.passed ? '✔' : '✘';
167
- const status = r.passed ? 'PASS' : 'FAIL';
168
- lines.push(` ${icon} [${status}] ${r.name} (${r.durationMs}ms)`);
169
- for (const f of r.failures) {
170
- lines.push(` → ${f}`);
171
- }
172
- }
173
-
174
- lines.push('');
175
- lines.push('───────────────────────────────────────────');
176
- lines.push(` Total: ${report.total} Passed: ${report.passed} Failed: ${report.failed} Duration: ${report.duration}ms`);
177
- lines.push('───────────────────────────────────────────');
178
- lines.push('');
179
-
180
- return lines.join('\n');
181
- }
1
+ /**
2
+ * Agent Testing Framework - Define test cases in OAD, run with `opc test`.
3
+ * Supports assertions on response content, tool calls, and latency.
4
+ */
5
+ import * as fs from 'fs';
6
+ import * as path from 'path';
7
+ import * as yaml from 'js-yaml';
8
+ import { AgentRuntime } from '../core/runtime';
9
+
10
+ export interface TestCase {
11
+ name: string;
12
+ input: string;
13
+ expect?: {
14
+ contains?: string[];
15
+ notContains?: string[];
16
+ toolCalled?: string[];
17
+ maxLatencyMs?: number;
18
+ };
19
+ }
20
+
21
+ export interface TestResult {
22
+ name: string;
23
+ passed: boolean;
24
+ durationMs: number;
25
+ response?: string;
26
+ failures: string[];
27
+ }
28
+
29
+ export interface TestReport {
30
+ total: number;
31
+ passed: number;
32
+ failed: number;
33
+ duration: number;
34
+ results: TestResult[];
35
+ }
36
+
37
+ /**
38
+ * Load test cases from OAD spec.testing or a separate test file.
39
+ */
40
+ export function loadTestCases(oadPath: string): TestCase[] {
41
+ const raw = fs.readFileSync(oadPath, 'utf-8');
42
+ const config = yaml.load(raw) as any;
43
+
44
+ // Check spec.testing.cases
45
+ if (config?.spec?.testing?.cases) {
46
+ return config.spec.testing.cases;
47
+ }
48
+
49
+ // Check for companion test file
50
+ const dir = path.dirname(oadPath);
51
+ const testFile = path.join(dir, 'tests.yaml');
52
+ if (fs.existsSync(testFile)) {
53
+ const testRaw = fs.readFileSync(testFile, 'utf-8');
54
+ const testConfig = yaml.load(testRaw) as any;
55
+ return testConfig?.cases ?? testConfig ?? [];
56
+ }
57
+
58
+ return [];
59
+ }
60
+
61
+ /**
62
+ * Run all test cases against an agent.
63
+ */
64
+ export async function runTests(oadPath: string): Promise<TestReport> {
65
+ const cases = loadTestCases(oadPath);
66
+ const results: TestResult[] = [];
67
+ const startTime = Date.now();
68
+
69
+ if (cases.length === 0) {
70
+ // Generate default smoke test
71
+ cases.push({
72
+ name: 'smoke-test',
73
+ input: 'Hello! What can you help me with?',
74
+ expect: { maxLatencyMs: 30000 },
75
+ });
76
+ }
77
+
78
+ const runtime = new AgentRuntime();
79
+ await runtime.loadConfig(oadPath);
80
+ const agent = await runtime.initialize();
81
+
82
+ for (const tc of cases) {
83
+ const result: TestResult = {
84
+ name: tc.name,
85
+ passed: true,
86
+ durationMs: 0,
87
+ failures: [],
88
+ };
89
+
90
+ const t0 = Date.now();
91
+ try {
92
+ const response = await agent.handleMessage({
93
+ id: `test_${Date.now()}`,
94
+ role: 'user',
95
+ content: tc.input,
96
+ timestamp: Date.now(),
97
+ });
98
+ result.durationMs = Date.now() - t0;
99
+ result.response = response.content;
100
+
101
+ if (tc.expect) {
102
+ // Check contains
103
+ if (tc.expect.contains) {
104
+ for (const s of tc.expect.contains) {
105
+ if (!response.content.toLowerCase().includes(s.toLowerCase())) {
106
+ result.failures.push(`Expected response to contain "${s}"`);
107
+ }
108
+ }
109
+ }
110
+ // Check notContains
111
+ if (tc.expect.notContains) {
112
+ for (const s of tc.expect.notContains) {
113
+ if (response.content.toLowerCase().includes(s.toLowerCase())) {
114
+ result.failures.push(`Expected response NOT to contain "${s}"`);
115
+ }
116
+ }
117
+ }
118
+ // Check latency
119
+ if (tc.expect.maxLatencyMs && result.durationMs > tc.expect.maxLatencyMs) {
120
+ result.failures.push(`Latency ${result.durationMs}ms exceeded max ${tc.expect.maxLatencyMs}ms`);
121
+ }
122
+ // Check tool calls (from metadata if available)
123
+ if (tc.expect.toolCalled && (response as any).toolsCalled) {
124
+ for (const tool of tc.expect.toolCalled) {
125
+ if (!(response as any).toolsCalled.includes(tool)) {
126
+ result.failures.push(`Expected tool "${tool}" to be called`);
127
+ }
128
+ }
129
+ }
130
+ }
131
+
132
+ result.passed = result.failures.length === 0;
133
+ } catch (err) {
134
+ result.durationMs = Date.now() - t0;
135
+ result.passed = false;
136
+ result.failures.push(`Error: ${err instanceof Error ? err.message : String(err)}`);
137
+ }
138
+
139
+ results.push(result);
140
+ }
141
+
142
+ const totalDuration = Date.now() - startTime;
143
+ const passed = results.filter(r => r.passed).length;
144
+
145
+ return {
146
+ total: results.length,
147
+ passed,
148
+ failed: results.length - passed,
149
+ duration: totalDuration,
150
+ results,
151
+ };
152
+ }
153
+
154
+ /**
155
+ * Format test report for console output.
156
+ */
157
+ export function formatReport(report: TestReport): string {
158
+ const lines: string[] = [];
159
+ lines.push('');
160
+ lines.push('═══════════════════════════════════════════');
161
+ lines.push(' OPC Agent Test Report');
162
+ lines.push('═══════════════════════════════════════════');
163
+ lines.push('');
164
+
165
+ for (const r of report.results) {
166
+ const icon = r.passed ? '✔' : '✘';
167
+ const status = r.passed ? 'PASS' : 'FAIL';
168
+ lines.push(` ${icon} [${status}] ${r.name} (${r.durationMs}ms)`);
169
+ for (const f of r.failures) {
170
+ lines.push(` → ${f}`);
171
+ }
172
+ }
173
+
174
+ lines.push('');
175
+ lines.push('───────────────────────────────────────────');
176
+ lines.push(` Total: ${report.total} Passed: ${report.passed} Failed: ${report.failed} Duration: ${report.duration}ms`);
177
+ lines.push('───────────────────────────────────────────');
178
+ lines.push('');
179
+
180
+ return lines.join('\n');
181
+ }
@@ -1,73 +1,73 @@
1
- import type { MCPTool, MCPToolResult } from './mcp';
2
-
3
- /**
4
- * Calculator Tool — v0.8.0
5
- * Safe math expression evaluation as an LLM function tool.
6
- */
7
- export const CalculatorTool: MCPTool = {
8
- name: 'calculator',
9
- description: 'Evaluate a mathematical expression. Supports basic arithmetic, powers, sqrt, abs, min, max, round, ceil, floor, PI, E.',
10
- inputSchema: {
11
- type: 'object',
12
- properties: {
13
- expression: {
14
- type: 'string',
15
- description: 'Mathematical expression to evaluate, e.g. "2 + 3 * 4" or "sqrt(144) + PI"',
16
- },
17
- },
18
- required: ['expression'],
19
- },
20
-
21
- async execute(input: Record<string, unknown>): Promise<MCPToolResult> {
22
- const expr = String(input.expression ?? '');
23
- try {
24
- const result = safeEval(expr);
25
- return { content: String(result) };
26
- } catch (err) {
27
- return { content: `Error: ${(err as Error).message}`, isError: true };
28
- }
29
- },
30
- };
31
-
32
- /** Safe math evaluator — no eval(), no arbitrary code */
33
- function safeEval(expr: string): number {
34
- // Whitelist: digits, operators, parens, dots, commas, spaces, and known functions
35
- const sanitized = expr.replace(/\s+/g, '');
36
- const allowed = /^[0-9+\-*/().,%^a-zA-Z_]+$/;
37
- if (!allowed.test(sanitized)) {
38
- throw new Error('Invalid characters in expression');
39
- }
40
-
41
- // Replace known math functions/constants
42
- const prepared = sanitized
43
- .replace(/\bPI\b/gi, String(Math.PI))
44
- .replace(/\bE\b/g, String(Math.E))
45
- .replace(/\bsqrt\b/gi, 'Math.sqrt')
46
- .replace(/\babs\b/gi, 'Math.abs')
47
- .replace(/\bmin\b/gi, 'Math.min')
48
- .replace(/\bmax\b/gi, 'Math.max')
49
- .replace(/\bround\b/gi, 'Math.round')
50
- .replace(/\bceil\b/gi, 'Math.ceil')
51
- .replace(/\bfloor\b/gi, 'Math.floor')
52
- .replace(/\bpow\b/gi, 'Math.pow')
53
- .replace(/\blog\b/gi, 'Math.log')
54
- .replace(/\blog10\b/gi, 'Math.log10')
55
- .replace(/\bsin\b/gi, 'Math.sin')
56
- .replace(/\bcos\b/gi, 'Math.cos')
57
- .replace(/\btan\b/gi, 'Math.tan')
58
- .replace(/\^/g, '**');
59
-
60
- // Block anything that isn't math
61
- if (/[a-zA-Z_]/.test(prepared.replace(/Math\.\w+/g, ''))) {
62
- throw new Error('Unsupported function or variable in expression');
63
- }
64
-
65
- // Use Function constructor with restricted scope
66
- const fn = new Function('Math', `"use strict"; return (${prepared});`);
67
- const result = fn(Math);
68
-
69
- if (typeof result !== 'number' || !isFinite(result)) {
70
- throw new Error('Expression did not evaluate to a finite number');
71
- }
72
- return result;
73
- }
1
+ import type { MCPTool, MCPToolResult } from './mcp';
2
+
3
+ /**
4
+ * Calculator Tool — v0.8.0
5
+ * Safe math expression evaluation as an LLM function tool.
6
+ */
7
+ export const CalculatorTool: MCPTool = {
8
+ name: 'calculator',
9
+ description: 'Evaluate a mathematical expression. Supports basic arithmetic, powers, sqrt, abs, min, max, round, ceil, floor, PI, E.',
10
+ inputSchema: {
11
+ type: 'object',
12
+ properties: {
13
+ expression: {
14
+ type: 'string',
15
+ description: 'Mathematical expression to evaluate, e.g. "2 + 3 * 4" or "sqrt(144) + PI"',
16
+ },
17
+ },
18
+ required: ['expression'],
19
+ },
20
+
21
+ async execute(input: Record<string, unknown>): Promise<MCPToolResult> {
22
+ const expr = String(input.expression ?? '');
23
+ try {
24
+ const result = safeEval(expr);
25
+ return { content: String(result) };
26
+ } catch (err) {
27
+ return { content: `Error: ${(err as Error).message}`, isError: true };
28
+ }
29
+ },
30
+ };
31
+
32
+ /** Safe math evaluator — no eval(), no arbitrary code */
33
+ function safeEval(expr: string): number {
34
+ // Whitelist: digits, operators, parens, dots, commas, spaces, and known functions
35
+ const sanitized = expr.replace(/\s+/g, '');
36
+ const allowed = /^[0-9+\-*/().,%^a-zA-Z_]+$/;
37
+ if (!allowed.test(sanitized)) {
38
+ throw new Error('Invalid characters in expression');
39
+ }
40
+
41
+ // Replace known math functions/constants
42
+ const prepared = sanitized
43
+ .replace(/\bPI\b/gi, String(Math.PI))
44
+ .replace(/\bE\b/g, String(Math.E))
45
+ .replace(/\bsqrt\b/gi, 'Math.sqrt')
46
+ .replace(/\babs\b/gi, 'Math.abs')
47
+ .replace(/\bmin\b/gi, 'Math.min')
48
+ .replace(/\bmax\b/gi, 'Math.max')
49
+ .replace(/\bround\b/gi, 'Math.round')
50
+ .replace(/\bceil\b/gi, 'Math.ceil')
51
+ .replace(/\bfloor\b/gi, 'Math.floor')
52
+ .replace(/\bpow\b/gi, 'Math.pow')
53
+ .replace(/\blog\b/gi, 'Math.log')
54
+ .replace(/\blog10\b/gi, 'Math.log10')
55
+ .replace(/\bsin\b/gi, 'Math.sin')
56
+ .replace(/\bcos\b/gi, 'Math.cos')
57
+ .replace(/\btan\b/gi, 'Math.tan')
58
+ .replace(/\^/g, '**');
59
+
60
+ // Block anything that isn't math
61
+ if (/[a-zA-Z_]/.test(prepared.replace(/Math\.\w+/g, ''))) {
62
+ throw new Error('Unsupported function or variable in expression');
63
+ }
64
+
65
+ // Use Function constructor with restricted scope
66
+ const fn = new Function('Math', `"use strict"; return (${prepared});`);
67
+ const result = fn(Math);
68
+
69
+ if (typeof result !== 'number' || !isFinite(result)) {
70
+ throw new Error('Expression did not evaluate to a finite number');
71
+ }
72
+ return result;
73
+ }