@polymorphism-tech/morph-spec 8.33.0 → 8.34.2

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 (176) hide show
  1. package/CLAUDE.md +3 -0
  2. package/README.md +2 -2
  3. package/bin/morph-spec.js +25 -1
  4. package/bin/validate.js +379 -379
  5. package/claude-plugin.json +1 -1
  6. package/docs/CHEATSHEET.md +1 -1
  7. package/docs/QUICKSTART.md +1 -1
  8. package/framework/CLAUDE.md +2 -2
  9. package/framework/MORPH.md +4 -5
  10. package/framework/agents.json +1 -0
  11. package/framework/docs/CLI.md +5 -2
  12. package/framework/hooks/README.md +11 -1
  13. package/framework/hooks/claude-code/core/block-destructive-worktree-removal.js +5 -8
  14. package/framework/hooks/claude-code/core/block-raw-worktree-add.js +79 -0
  15. package/framework/hooks/claude-code/core/chain.js +27 -7
  16. package/framework/hooks/claude-code/core/worktree-context.js +83 -14
  17. package/framework/hooks/claude-code/core/worktree-create.js +206 -0
  18. package/framework/hooks/claude-code/core/worktree-env-inject.js +87 -0
  19. package/framework/hooks/claude-code/core/worktree-remove.js +81 -0
  20. package/framework/hooks/claude-code/statusline.sh +11 -11
  21. package/framework/hooks/git/commit-msg/conventional-commits.sh +33 -33
  22. package/framework/hooks/git/pre-commit/agents.sh +25 -25
  23. package/framework/hooks/git/pre-commit/orchestrator.sh +64 -64
  24. package/framework/hooks/git/pre-commit/specs.sh +50 -50
  25. package/framework/hooks/git/pre-push/run-tests.sh +44 -44
  26. package/framework/hooks/shared/command-tokens.js +23 -0
  27. package/framework/hooks/shared/e2e-check.js +0 -0
  28. package/framework/hooks/shared/git-paths.js +25 -0
  29. package/framework/hooks/shared/hook-response.js +17 -0
  30. package/framework/hooks/shared/morph-cli.js +134 -0
  31. package/framework/hooks/shared/stdin-reader.js +26 -26
  32. package/framework/rules/infrastructure-standards.md +3 -1
  33. package/framework/skills/core/create-umbraco-lp/SKILL.md +1 -1
  34. package/framework/skills/core/migrate-lp-to-umbraco/SKILL.md +1 -1
  35. package/framework/skills/core/morph-apply/SKILL.md +3 -3
  36. package/framework/skills/core/morph-brainstorming/SKILL.md +1 -1
  37. package/framework/skills/core/morph-eval/SKILL.md +1 -1
  38. package/framework/skills/core/morph-hotfix/SKILL.md +1 -1
  39. package/framework/skills/core/morph-implement/SKILL.md +3 -1
  40. package/framework/skills/core/morph-plan/SKILL.md +6 -3
  41. package/framework/skills/core/morph-proposal/SKILL.md +1 -1
  42. package/framework/skills/core/morph-review/SKILL.md +1 -1
  43. package/framework/skills/core/morph-stack-scan/SKILL.md +1 -1
  44. package/framework/skills/core/morph-standards/SKILL.md +1 -1
  45. package/framework/skills/core/morph-uiux/SKILL.md +1 -1
  46. package/framework/standards/STANDARDS.json +30 -0
  47. package/framework/standards/ai-agents/agent-session.md +160 -160
  48. package/framework/standards/ai-agents/batching.md +193 -193
  49. package/framework/standards/ai-agents/context-providers.md +188 -188
  50. package/framework/standards/ai-agents/mcp-server.md +250 -250
  51. package/framework/standards/ai-agents/mcp-tools.md +197 -197
  52. package/framework/standards/ai-agents/middleware-patterns.md +207 -207
  53. package/framework/standards/ai-agents/modalities-image-gen.md +173 -173
  54. package/framework/standards/ai-agents/modalities-speech-to-text.md +171 -171
  55. package/framework/standards/ai-agents/modalities-text-to-speech.md +181 -181
  56. package/framework/standards/ai-agents/modalities-vision.md +201 -201
  57. package/framework/standards/ai-agents/multi-agent-patterns.md +155 -155
  58. package/framework/standards/ai-agents/production.md +231 -231
  59. package/framework/standards/ai-agents/prompts-in-db.md +214 -214
  60. package/framework/standards/ai-agents/providers/model-registry.md +158 -158
  61. package/framework/standards/ai-agents/service-tier-flex.md +135 -135
  62. package/framework/standards/ai-agents/structured-output.md +249 -249
  63. package/framework/standards/ai-agents/sweet-spot.md +109 -109
  64. package/framework/standards/infrastructure/docker/local-compose-isolation.md +88 -0
  65. package/framework/templates/IDE-SNIPPETS.md +262 -262
  66. package/framework/templates/REGISTRY.json +1 -1
  67. package/framework/templates/code/dotnet/abstractions/api-endpoint.cs +7 -7
  68. package/framework/templates/code/dotnet/ai-agents/AgentBootstrap.cs.template +39 -39
  69. package/framework/templates/code/dotnet/ai-agents/EmbeddingService.cs.template +68 -68
  70. package/framework/templates/code/dotnet/ai-agents/ModelRegistry.cs.template +117 -117
  71. package/framework/templates/code/dotnet/ai-agents/PromptRepository.cs.template +95 -95
  72. package/framework/templates/code/dotnet/ai-agents/ResilientToolMiddleware.cs.template +61 -61
  73. package/framework/templates/code/dotnet/ai-agents/VectorSearchTool.cs.template +107 -107
  74. package/framework/templates/code/dotnet/ai-agents/model-registry.json.template +58 -58
  75. package/framework/templates/code/dotnet/contracts/contracts-vsa.cs +281 -281
  76. package/framework/templates/code/dotnet/database/migration.cs +83 -83
  77. package/framework/templates/code/dotnet/feature/endpoint.cs.hbs +21 -21
  78. package/framework/templates/code/dotnet/feature/request.cs.hbs +4 -4
  79. package/framework/templates/code/dotnet/feature/response.cs.hbs +4 -4
  80. package/framework/templates/code/dotnet/feature/validator.cs.hbs +15 -15
  81. package/framework/templates/code/sql/ai-agents/agent-prompts-migration.sql.template +58 -58
  82. package/framework/templates/code/sql/ai-agents/pgvector-embeddings-migration.sql.template +91 -91
  83. package/framework/templates/code/typescript/contracts.ts +167 -167
  84. package/framework/templates/docs/clarifications.md +253 -253
  85. package/framework/templates/docs/schema-analysis.md +119 -119
  86. package/framework/templates/docs/ui-components.md +52 -52
  87. package/framework/templates/docs/ui-design-system.md +76 -76
  88. package/framework/templates/docs/ui-flows.md +167 -167
  89. package/framework/templates/docs/ui-mockups.md +98 -98
  90. package/framework/templates/feature/decisions.md +122 -122
  91. package/framework/templates/frontend/nextjs/Dockerfile.nextjs.hbs +43 -43
  92. package/framework/templates/frontend/nextjs/client-component.tsx.hbs +26 -26
  93. package/framework/templates/frontend/nextjs/env.mjs.hbs +32 -32
  94. package/framework/templates/frontend/nextjs/feature-form.tsx.hbs +56 -56
  95. package/framework/templates/frontend/nextjs/page.tsx.hbs +22 -22
  96. package/framework/templates/frontend/nextjs/tsconfig.json.hbs +26 -26
  97. package/framework/templates/frontend/nextjs/use-feature.ts.hbs +54 -54
  98. package/framework/templates/infrastructure/docker/Dockerfile.template +58 -58
  99. package/framework/templates/infrastructure/docker/docker-compose.template.yml +24 -14
  100. package/framework/templates/infrastructure/docker/dockerfile-api.dockerfile +38 -38
  101. package/framework/templates/infrastructure/docker/dockerfile-web.dockerfile +48 -48
  102. package/framework/templates/infrastructure/github/README.md +56 -56
  103. package/framework/templates/infrastructure/github/actions/docker-build-push/action.yml.hbs +45 -45
  104. package/framework/templates/infrastructure/github/actions/health-check/action.yml.hbs +27 -27
  105. package/framework/templates/infrastructure/github/workflows/docker-build-push.yml.hbs +59 -59
  106. package/framework/templates/infrastructure/github/workflows/dotnet-build.yml.hbs +39 -39
  107. package/framework/templates/integrations/asaas-client.cs +387 -387
  108. package/framework/templates/integrations/asaas-webhook.cs +351 -351
  109. package/framework/templates/meta-prompts/fusion/fusion-agent.md +76 -76
  110. package/framework/templates/meta-prompts/fusion/fusion-aggregator.md +100 -100
  111. package/framework/templates/meta-prompts/validators/pre-commit-validator.md +95 -95
  112. package/framework/templates/project/validate.js +120 -120
  113. package/framework/templates/state.template.json +17 -17
  114. package/framework/templates/ui/design-system.css +226 -226
  115. package/package.json +1 -1
  116. package/src/commands/dashboard/dashboard.js +53 -53
  117. package/src/commands/e2e/e2e.js +77 -39
  118. package/src/commands/index.js +13 -13
  119. package/src/commands/project/checkpoint.js +209 -209
  120. package/src/commands/project/doctor.js +8 -6
  121. package/src/commands/project/env.js +95 -0
  122. package/src/commands/project/worktree.js +266 -52
  123. package/src/commands/state/delete.js +8 -0
  124. package/src/commands/state/finish.js +29 -34
  125. package/src/core/state/phase-state-machine.js +197 -197
  126. package/src/core/templates/template-renderer.js +477 -477
  127. package/src/dashboard/app/components/DetailPanel.jsx +11 -11
  128. package/src/dashboard/app/components/FlowGraph.jsx +17 -17
  129. package/src/dashboard/app/components/SearchFilter.jsx +22 -22
  130. package/src/dashboard/app/hooks/useDetailPanel.js +15 -15
  131. package/src/dashboard/app/main.jsx +13 -13
  132. package/src/dashboard/app/pages/AgentMap.jsx +370 -370
  133. package/src/dashboard/app/pages/HookInspector.jsx +430 -418
  134. package/src/dashboard/app/pages/PhasePipeline.jsx +408 -408
  135. package/src/dashboard/app/pages/RulesMap.jsx +273 -273
  136. package/src/dashboard/app/pages/SkillsLibrary.jsx +366 -366
  137. package/src/dashboard/app/pages/Standards.jsx +332 -332
  138. package/src/dashboard/app/styles/index.css +47 -47
  139. package/src/dashboard/index.html +12 -12
  140. package/src/dashboard/package.json +24 -24
  141. package/src/dashboard/scanner/hooks-scanner.js +5 -1
  142. package/src/dashboard/scanner/phases-scanner.js +15 -15
  143. package/src/dashboard/scanner/rubrics-scanner.js +12 -12
  144. package/src/dashboard/scanner/rules-scanner.js +37 -37
  145. package/src/dashboard/scanner/skills-scanner.js +43 -43
  146. package/src/dashboard/scanner/squads-scanner.js +13 -13
  147. package/src/dashboard/vite-plugin-morph-data.js +17 -17
  148. package/src/dashboard/vite.config.js +24 -24
  149. package/src/lib/detectors/design-system-detector.js +188 -188
  150. package/src/lib/detectors/index.js +217 -217
  151. package/src/lib/e2e/compose-env.js +6 -11
  152. package/src/lib/e2e/stack-teardown.js +115 -0
  153. package/src/lib/generators/architecture-scanner.js +2 -1
  154. package/src/lib/generators/design-system-generator.js +508 -508
  155. package/src/lib/generators/settings-generator.js +209 -209
  156. package/src/lib/graph/graphifyignore.js +2 -1
  157. package/src/lib/hooks/hook-executor.js +167 -167
  158. package/src/lib/installers/mcp-installer.js +719 -719
  159. package/src/lib/tasks/task-classifier.js +30 -30
  160. package/src/lib/tasks/task-parser.js +209 -209
  161. package/src/lib/tasks/test-runner.js +21 -4
  162. package/src/lib/validators/checkpoint-persistence.js +39 -39
  163. package/src/lib/validators/css/css-validator.js +340 -340
  164. package/src/lib/validators/shared/emit-validator-dispatch.js +14 -14
  165. package/src/lib/validators/shared/issue-counter.js +14 -14
  166. package/src/lib/validators/verify-runner.js +7 -0
  167. package/src/lib/worktree/dirs.js +32 -0
  168. package/src/lib/worktree/env-allocator.js +45 -11
  169. package/src/lib/worktree/worktree-audit.js +40 -2
  170. package/src/utils/agent-id-mapper.js +11 -11
  171. package/src/utils/claude-settings-manager.js +300 -300
  172. package/src/utils/file-copier.js +18 -13
  173. package/src/utils/hooks-installer.js +46 -3
  174. package/src/utils/skills-installer.js +187 -187
  175. package/framework/templates/.idea/morph-templates.xml +0 -92
  176. package/framework/templates/.vscode/morph-templates.code-snippets +0 -186
@@ -1,231 +1,231 @@
1
- # Microsoft Agent Framework — Production Patterns
2
-
3
- > **Scope:** stacks=["dotnet"]
4
- > **Layer:** 1 (on-keyword)
5
- > **Keywords:** middleware, observability, telemetry, mcp, a2a, caching, anti-hallucination, production agent
6
- > **Read by Claude in:** implement (when wiring middleware/telemetry/MCP into an agent)
7
-
8
- **Verified against:** MAF 1.0 GA. Last-verified: 2026-05-19.
9
-
10
- ---
11
-
12
- ## Middleware Pipeline
13
-
14
- Três tipos de middleware interceptam o comportamento do agent em camadas distintas:
15
-
16
- | Tipo | Intercepts | Use Case |
17
- |------|-----------|----------|
18
- | **Agent Run** | Invocação completa do agent | Logging, auth, rate limiting |
19
- | **Function Calling** | Chamadas individuais de tools | Error handling, validação, auditoria |
20
- | **IChatClient** | Chamadas brutas ao LLM | Telemetry, caching, content filtering |
21
-
22
- O **tratamento canônico** de middleware — código dos 3 níveis, registro via `.AsBuilder().Use(...)`,
23
- e os patterns de telemetria, resilência (retry + rate-limit) e anti-hallucination — está em
24
- **`ai-agents-middleware-patterns`**. Não duplicado aqui.
25
-
26
- > Em produção: registre `ResilientToolMiddleware` (ou equivalente) em qualquer agent que chame
27
- > APIs externas, e um middleware de telemetria no `IChatClient`. Ver `ai-agents-middleware-patterns`
28
- > para o código e a decision tree de qual nível usar.
29
-
30
- ---
31
-
32
- ## A2A Protocol (Agent-to-Agent)
33
-
34
- Exponha agents como serviços web interoperáveis via protocolo A2A padronizado.
35
-
36
- **Package (GA):**
37
-
38
- ```xml
39
- <PackageReference Include="Microsoft.Agents.AI.A2A" Version="1.0.*" />
40
- ```
41
-
42
- ### Expor um agent via A2A
43
-
44
- ```csharp
45
- // Program.cs
46
- using Microsoft.Agents.AI;
47
- using Microsoft.Agents.AI.A2A;
48
- using Microsoft.Extensions.AI;
49
-
50
- var builder = WebApplication.CreateBuilder(args);
51
-
52
- // Registrar agent via Model Registry
53
- IChatClient chatClient = /* from ModelRegistry singleton */;
54
- builder.Services.AddSingleton(chatClient);
55
-
56
- builder.AddAIAgent("math", instructions: "You are a math expert.");
57
- builder.AddAIAgent("history", instructions: "You are a history expert.");
58
-
59
- var app = builder.Build();
60
-
61
- app.MapA2A("math", path: "/a2a/math", agentCard: new()
62
- {
63
- Name = "Math Agent",
64
- Description = "Expert in mathematics and calculations.",
65
- Version = "1.0"
66
- });
67
-
68
- app.MapA2A("history", path: "/a2a/history", agentCard: new()
69
- {
70
- Name = "History Agent",
71
- Description = "Expert in world history.",
72
- Version = "1.0"
73
- });
74
-
75
- app.Run();
76
- ```
77
-
78
- ### Consumir um agent A2A remoto
79
-
80
- ```csharp
81
- using Microsoft.Agents.AI.A2A;
82
-
83
- // Wrap a remote A2A endpoint as a local AIAgent
84
- AIAgent remoteAgent = new A2AClient(new Uri("https://agents.example.com/a2a/math"))
85
- .AsAIAgent();
86
-
87
- var response = await remoteAgent.RunAsync("What is the integral of x^2?");
88
- ```
89
-
90
- ### Agent Card Discovery
91
-
92
- ```
93
- GET /a2a/math/v1/card → metadata (name, description, capabilities)
94
- POST /a2a/math/v1/message:stream → send message, get streaming response
95
- ```
96
-
97
- | Feature A2A | Descrição |
98
- |-------------|-----------|
99
- | Agent Discovery | Agent Cards com metadados padronizados |
100
- | Cross-Framework | Agent .NET pode conversar com agent Python |
101
- | Context Persistence | `contextId` mantém conversa entre mensagens |
102
- | Streaming | Server-Sent Events para respostas em tempo real |
103
-
104
- ---
105
-
106
- ## MCP Integration (Model Context Protocol)
107
-
108
- Agents MAF podem consumir tools de **MCP servers externos** (GitHub, filesystem, Postgres,
109
- servidores internos) e expor capacidades do projeto **como** um MCP server.
110
-
111
- Esses dois casos têm tratamento canônico em standards dedicados — **não duplicado aqui**:
112
-
113
- | Caso | Standard |
114
- |------|----------|
115
- | Agente MAF consome um MCP server externo | `ai-agents-mcp-tools` (`McpClientFactory.CreateAsync`, transports, `await using`) |
116
- | Projeto expõe capacidades como MCP server | `ai-agents-mcp-server` (`[McpServerTool]`, `AddMcpServer`, stdio/HTTP) |
117
-
118
- > Em produção, lembre: MCP client sempre com `await using` (dispose do transporte); MCP server
119
- > exposto via HTTP exige autenticação. Detalhes nos standards acima.
120
-
121
- ---
122
-
123
- ## Caching Patterns
124
-
125
- ### Semantic Caching (Redis)
126
-
127
- Cache de respostas LLM por similaridade semântica — evita inferências repetidas.
128
-
129
- > **Illustrative pseudo-code** — the exact Redis vector-search calls depend on your Redis client. For a complete, working semantic-search implementation see `ai-agents-rag-custom-pgvector` (Onda 2). The point of this section is the *pattern*: embed the query, vector-search a cache, return on a similarity hit.
130
-
131
- ```csharp
132
- public async Task<string?> GetCachedResponseAsync(string question)
133
- {
134
- var questionEmbedding = await _embedding.GenerateAsync(question);
135
-
136
- // pseudo-code: perform a KNN vector search against the cache index
137
- var (similarityScore, cachedResponse) = await VectorSearchCacheAsync(
138
- "idx:semantic_cache", questionEmbedding);
139
-
140
- if (similarityScore >= 0.95) // cosine similarity: higher = more similar
141
- return cachedResponse;
142
-
143
- return null; // cache miss
144
- }
145
- ```
146
-
147
- **Cost savings:** embedding ~$0.0001/1K tokens vs inferência ~$0.01-0.03/1K tokens = **100-300x savings**.
148
-
149
- ### Hybrid Cache (L1 + L2)
150
-
151
- ```csharp
152
- builder.Services.AddHybridCache(options =>
153
- {
154
- options.DefaultEntryOptions = new HybridCacheEntryOptions
155
- {
156
- Expiration = TimeSpan.FromMinutes(10),
157
- LocalCacheExpiration = TimeSpan.FromMinutes(1)
158
- };
159
- });
160
-
161
- // Uso num tool
162
- [Description("Gets weather for a location")]
163
- public async Task<WeatherResponse> GetWeatherAsync(
164
- string location, string date, CancellationToken ct)
165
- {
166
- return await _cache.GetOrCreateAsync(
167
- $"weather:{location}:{date}",
168
- async token => await FetchFromApiAsync(location, date, token),
169
- cancellationToken: ct);
170
- }
171
- ```
172
-
173
- | Level | Location | Speed | Scope |
174
- |-------|----------|-------|-------|
175
- | L1 | Process memory | Fastest | Per instance |
176
- | L2 | Redis | Very fast | Distributed |
177
-
178
- ---
179
-
180
- ## Observability / OpenTelemetry
181
-
182
- Setup, enrichment (span attributes, correlation IDs, structured logging) e o workflow de investigação/análise de logs em produção têm **tratamento canônico** em **`ai-agents-observability-patterns`**. Não duplicado aqui.
183
-
184
- > Em produção: source name sempre `AgentOpenTelemetryConsts.DefaultSourceName`; `EnableSensitiveTelemetryData = false` fora de dev/staging. Ver `ai-agents-observability-patterns` para o setup completo, o que preencher em cada span/log, e o workflow de investigação quando um incidente for reportado.
185
-
186
- ---
187
-
188
- ## Security Checklist (verifiable by morph-eval)
189
-
190
- - [ ] API keys nunca em `appsettings.json` no repositório — bind de environment variables ou secret store
191
- - [ ] Variáveis de provider (`OpenAI__ApiKey`, `Google__ApiKey`, `Anthropic__ApiKey`) em env vars / Docker secrets
192
- - [ ] Input validation middleware registrado em agents públicos (PII detection, injection prevention)
193
- - [ ] Output filtering / guardrails para conteúdo prejudicial
194
- - [ ] Rate limiting middleware em agents expostos publicamente
195
- - [ ] Endpoints A2A protegidos com autenticação
196
- - [ ] Structured output usado para respostas — evita parsing de texto livre
197
- - [ ] OpenTelemetry com `AgentOpenTelemetryConsts.DefaultSourceName` configurado
198
-
199
- ---
200
-
201
- ## Production Checklist (verifiable by morph-eval)
202
-
203
- - [ ] `Microsoft.Agents.AI` 1.0.* no `.csproj`; nenhum pacote `Experimental.*` ou SK
204
- - [ ] Middleware registrado via `.AsBuilder().Use(...)` antes de `Build()`
205
- - [ ] `ResilientToolMiddleware` (ou equivalente) em agents que chamam APIs externas
206
- - [ ] Semantic cache configurado se o agent recebe queries repetitivas
207
- - [ ] OpenTelemetry source = `AgentOpenTelemetryConsts.DefaultSourceName`
208
- - [ ] Dados sensíveis de telemetry desabilitados em produção (`EnableSensitiveTelemetryData = false`)
209
- - [ ] A2A endpoints com `MapA2A` protegidos por auth se expostos na internet
210
- - [ ] MCP client `await using` para dispose correto do transporte
211
-
212
- ---
213
-
214
- ## References
215
-
216
- - `ai-agents-setup` — pacotes, DI, quick start
217
- - `ai-agents-sweet-spot` — decisão de arquitetura antes de middleware
218
- - `ai-agents-providers-model-registry` — Model Registry para `_modelRegistry.GetChatClient(...)`
219
- - `ai-agents-middleware-patterns` — **tratamento canônico** de middleware (código dos 3 níveis, telemetria, resilência, anti-hallucination)
220
- - `ai-agents-observability-patterns` — **tratamento canônico** de observability (setup OpenTelemetry, enrichment de spans/logs, workflow de investigação)
221
- - `ai-agents-mcp-tools` — agente MAF consumindo MCP servers externos
222
- - `ai-agents-mcp-server` — projeto expondo capacidades como MCP server
223
- - `ai-agents-rag-custom-pgvector` — implementação completa de busca semântica (referida em Caching)
224
- - [Agent Middleware](https://learn.microsoft.com/agent-framework/user-guide/agents/agent-middleware)
225
- - [A2A Integration](https://learn.microsoft.com/agent-framework/user-guide/hosting/agent-to-agent-integration)
226
- - [Using MCP Tools](https://learn.microsoft.com/agent-framework/user-guide/model-context-protocol/using-mcp-tools)
227
- - [MCP C# SDK](https://github.com/modelcontextprotocol/csharp-sdk)
228
-
229
- ---
230
-
231
- *MORPH-SPEC by Polymorphism Tech — ai-agents/production.md v2.1 (2026-05-19)*
1
+ # Microsoft Agent Framework — Production Patterns
2
+
3
+ > **Scope:** stacks=["dotnet"]
4
+ > **Layer:** 1 (on-keyword)
5
+ > **Keywords:** middleware, observability, telemetry, mcp, a2a, caching, anti-hallucination, production agent
6
+ > **Read by Claude in:** implement (when wiring middleware/telemetry/MCP into an agent)
7
+
8
+ **Verified against:** MAF 1.0 GA. Last-verified: 2026-05-19.
9
+
10
+ ---
11
+
12
+ ## Middleware Pipeline
13
+
14
+ Três tipos de middleware interceptam o comportamento do agent em camadas distintas:
15
+
16
+ | Tipo | Intercepts | Use Case |
17
+ |------|-----------|----------|
18
+ | **Agent Run** | Invocação completa do agent | Logging, auth, rate limiting |
19
+ | **Function Calling** | Chamadas individuais de tools | Error handling, validação, auditoria |
20
+ | **IChatClient** | Chamadas brutas ao LLM | Telemetry, caching, content filtering |
21
+
22
+ O **tratamento canônico** de middleware — código dos 3 níveis, registro via `.AsBuilder().Use(...)`,
23
+ e os patterns de telemetria, resilência (retry + rate-limit) e anti-hallucination — está em
24
+ **`ai-agents-middleware-patterns`**. Não duplicado aqui.
25
+
26
+ > Em produção: registre `ResilientToolMiddleware` (ou equivalente) em qualquer agent que chame
27
+ > APIs externas, e um middleware de telemetria no `IChatClient`. Ver `ai-agents-middleware-patterns`
28
+ > para o código e a decision tree de qual nível usar.
29
+
30
+ ---
31
+
32
+ ## A2A Protocol (Agent-to-Agent)
33
+
34
+ Exponha agents como serviços web interoperáveis via protocolo A2A padronizado.
35
+
36
+ **Package (GA):**
37
+
38
+ ```xml
39
+ <PackageReference Include="Microsoft.Agents.AI.A2A" Version="1.0.*" />
40
+ ```
41
+
42
+ ### Expor um agent via A2A
43
+
44
+ ```csharp
45
+ // Program.cs
46
+ using Microsoft.Agents.AI;
47
+ using Microsoft.Agents.AI.A2A;
48
+ using Microsoft.Extensions.AI;
49
+
50
+ var builder = WebApplication.CreateBuilder(args);
51
+
52
+ // Registrar agent via Model Registry
53
+ IChatClient chatClient = /* from ModelRegistry singleton */;
54
+ builder.Services.AddSingleton(chatClient);
55
+
56
+ builder.AddAIAgent("math", instructions: "You are a math expert.");
57
+ builder.AddAIAgent("history", instructions: "You are a history expert.");
58
+
59
+ var app = builder.Build();
60
+
61
+ app.MapA2A("math", path: "/a2a/math", agentCard: new()
62
+ {
63
+ Name = "Math Agent",
64
+ Description = "Expert in mathematics and calculations.",
65
+ Version = "1.0"
66
+ });
67
+
68
+ app.MapA2A("history", path: "/a2a/history", agentCard: new()
69
+ {
70
+ Name = "History Agent",
71
+ Description = "Expert in world history.",
72
+ Version = "1.0"
73
+ });
74
+
75
+ app.Run();
76
+ ```
77
+
78
+ ### Consumir um agent A2A remoto
79
+
80
+ ```csharp
81
+ using Microsoft.Agents.AI.A2A;
82
+
83
+ // Wrap a remote A2A endpoint as a local AIAgent
84
+ AIAgent remoteAgent = new A2AClient(new Uri("https://agents.example.com/a2a/math"))
85
+ .AsAIAgent();
86
+
87
+ var response = await remoteAgent.RunAsync("What is the integral of x^2?");
88
+ ```
89
+
90
+ ### Agent Card Discovery
91
+
92
+ ```
93
+ GET /a2a/math/v1/card → metadata (name, description, capabilities)
94
+ POST /a2a/math/v1/message:stream → send message, get streaming response
95
+ ```
96
+
97
+ | Feature A2A | Descrição |
98
+ |-------------|-----------|
99
+ | Agent Discovery | Agent Cards com metadados padronizados |
100
+ | Cross-Framework | Agent .NET pode conversar com agent Python |
101
+ | Context Persistence | `contextId` mantém conversa entre mensagens |
102
+ | Streaming | Server-Sent Events para respostas em tempo real |
103
+
104
+ ---
105
+
106
+ ## MCP Integration (Model Context Protocol)
107
+
108
+ Agents MAF podem consumir tools de **MCP servers externos** (GitHub, filesystem, Postgres,
109
+ servidores internos) e expor capacidades do projeto **como** um MCP server.
110
+
111
+ Esses dois casos têm tratamento canônico em standards dedicados — **não duplicado aqui**:
112
+
113
+ | Caso | Standard |
114
+ |------|----------|
115
+ | Agente MAF consome um MCP server externo | `ai-agents-mcp-tools` (`McpClientFactory.CreateAsync`, transports, `await using`) |
116
+ | Projeto expõe capacidades como MCP server | `ai-agents-mcp-server` (`[McpServerTool]`, `AddMcpServer`, stdio/HTTP) |
117
+
118
+ > Em produção, lembre: MCP client sempre com `await using` (dispose do transporte); MCP server
119
+ > exposto via HTTP exige autenticação. Detalhes nos standards acima.
120
+
121
+ ---
122
+
123
+ ## Caching Patterns
124
+
125
+ ### Semantic Caching (Redis)
126
+
127
+ Cache de respostas LLM por similaridade semântica — evita inferências repetidas.
128
+
129
+ > **Illustrative pseudo-code** — the exact Redis vector-search calls depend on your Redis client. For a complete, working semantic-search implementation see `ai-agents-rag-custom-pgvector` (Onda 2). The point of this section is the *pattern*: embed the query, vector-search a cache, return on a similarity hit.
130
+
131
+ ```csharp
132
+ public async Task<string?> GetCachedResponseAsync(string question)
133
+ {
134
+ var questionEmbedding = await _embedding.GenerateAsync(question);
135
+
136
+ // pseudo-code: perform a KNN vector search against the cache index
137
+ var (similarityScore, cachedResponse) = await VectorSearchCacheAsync(
138
+ "idx:semantic_cache", questionEmbedding);
139
+
140
+ if (similarityScore >= 0.95) // cosine similarity: higher = more similar
141
+ return cachedResponse;
142
+
143
+ return null; // cache miss
144
+ }
145
+ ```
146
+
147
+ **Cost savings:** embedding ~$0.0001/1K tokens vs inferência ~$0.01-0.03/1K tokens = **100-300x savings**.
148
+
149
+ ### Hybrid Cache (L1 + L2)
150
+
151
+ ```csharp
152
+ builder.Services.AddHybridCache(options =>
153
+ {
154
+ options.DefaultEntryOptions = new HybridCacheEntryOptions
155
+ {
156
+ Expiration = TimeSpan.FromMinutes(10),
157
+ LocalCacheExpiration = TimeSpan.FromMinutes(1)
158
+ };
159
+ });
160
+
161
+ // Uso num tool
162
+ [Description("Gets weather for a location")]
163
+ public async Task<WeatherResponse> GetWeatherAsync(
164
+ string location, string date, CancellationToken ct)
165
+ {
166
+ return await _cache.GetOrCreateAsync(
167
+ $"weather:{location}:{date}",
168
+ async token => await FetchFromApiAsync(location, date, token),
169
+ cancellationToken: ct);
170
+ }
171
+ ```
172
+
173
+ | Level | Location | Speed | Scope |
174
+ |-------|----------|-------|-------|
175
+ | L1 | Process memory | Fastest | Per instance |
176
+ | L2 | Redis | Very fast | Distributed |
177
+
178
+ ---
179
+
180
+ ## Observability / OpenTelemetry
181
+
182
+ Setup, enrichment (span attributes, correlation IDs, structured logging) e o workflow de investigação/análise de logs em produção têm **tratamento canônico** em **`ai-agents-observability-patterns`**. Não duplicado aqui.
183
+
184
+ > Em produção: source name sempre `AgentOpenTelemetryConsts.DefaultSourceName`; `EnableSensitiveTelemetryData = false` fora de dev/staging. Ver `ai-agents-observability-patterns` para o setup completo, o que preencher em cada span/log, e o workflow de investigação quando um incidente for reportado.
185
+
186
+ ---
187
+
188
+ ## Security Checklist (verifiable by morph-eval)
189
+
190
+ - [ ] API keys nunca em `appsettings.json` no repositório — bind de environment variables ou secret store
191
+ - [ ] Variáveis de provider (`OpenAI__ApiKey`, `Google__ApiKey`, `Anthropic__ApiKey`) em env vars / Docker secrets
192
+ - [ ] Input validation middleware registrado em agents públicos (PII detection, injection prevention)
193
+ - [ ] Output filtering / guardrails para conteúdo prejudicial
194
+ - [ ] Rate limiting middleware em agents expostos publicamente
195
+ - [ ] Endpoints A2A protegidos com autenticação
196
+ - [ ] Structured output usado para respostas — evita parsing de texto livre
197
+ - [ ] OpenTelemetry com `AgentOpenTelemetryConsts.DefaultSourceName` configurado
198
+
199
+ ---
200
+
201
+ ## Production Checklist (verifiable by morph-eval)
202
+
203
+ - [ ] `Microsoft.Agents.AI` 1.0.* no `.csproj`; nenhum pacote `Experimental.*` ou SK
204
+ - [ ] Middleware registrado via `.AsBuilder().Use(...)` antes de `Build()`
205
+ - [ ] `ResilientToolMiddleware` (ou equivalente) em agents que chamam APIs externas
206
+ - [ ] Semantic cache configurado se o agent recebe queries repetitivas
207
+ - [ ] OpenTelemetry source = `AgentOpenTelemetryConsts.DefaultSourceName`
208
+ - [ ] Dados sensíveis de telemetry desabilitados em produção (`EnableSensitiveTelemetryData = false`)
209
+ - [ ] A2A endpoints com `MapA2A` protegidos por auth se expostos na internet
210
+ - [ ] MCP client `await using` para dispose correto do transporte
211
+
212
+ ---
213
+
214
+ ## References
215
+
216
+ - `ai-agents-setup` — pacotes, DI, quick start
217
+ - `ai-agents-sweet-spot` — decisão de arquitetura antes de middleware
218
+ - `ai-agents-providers-model-registry` — Model Registry para `_modelRegistry.GetChatClient(...)`
219
+ - `ai-agents-middleware-patterns` — **tratamento canônico** de middleware (código dos 3 níveis, telemetria, resilência, anti-hallucination)
220
+ - `ai-agents-observability-patterns` — **tratamento canônico** de observability (setup OpenTelemetry, enrichment de spans/logs, workflow de investigação)
221
+ - `ai-agents-mcp-tools` — agente MAF consumindo MCP servers externos
222
+ - `ai-agents-mcp-server` — projeto expondo capacidades como MCP server
223
+ - `ai-agents-rag-custom-pgvector` — implementação completa de busca semântica (referida em Caching)
224
+ - [Agent Middleware](https://learn.microsoft.com/agent-framework/user-guide/agents/agent-middleware)
225
+ - [A2A Integration](https://learn.microsoft.com/agent-framework/user-guide/hosting/agent-to-agent-integration)
226
+ - [Using MCP Tools](https://learn.microsoft.com/agent-framework/user-guide/model-context-protocol/using-mcp-tools)
227
+ - [MCP C# SDK](https://github.com/modelcontextprotocol/csharp-sdk)
228
+
229
+ ---
230
+
231
+ *MORPH-SPEC by Polymorphism Tech — ai-agents/production.md v2.1 (2026-05-19)*