[aw] PR Code Quality Reviewer failed

[github-actions]

Workflow Failure

Workflow: PR Code Quality Reviewer
Branch: copilot/fix-stdio-container-env-secrets
Run: https://github.com/github/gh-aw/actions/runs/30683628677
Pull Request: #49450

Warning

Engine Failure: The copilot engine terminated unexpectedly.

Last agent output:

{"type":"tool.execution_complete","timestamp":"2026-08-01T06:25:50.976Z","data":{"toolName":"bash","mcpServerName":"","success":true,"result":{"content":"{\"additions\":127,\"body\":\"For non-Copilot engines (Claude, Codex, etc.), secrets in stdio/container custom MCP server `env:` blocks were substituted as bare `${VAR}` inside an unquoted heredoc. Bash expands those before the MCP gateway parses the config, so any secret containing `\\\"` or `\\\\` silently corrupts the JSON and the **Start MCP Gateway** step fails.\\n\\n## Root cause\\n\\n`renderCustomMCPEnvVars` had separate branches for Copilot (`\\\\${VAR}`, heredoc-safe) and non-Copilot (`${VAR}`, bash-expanded). The Copilot path was correct; the non-Copilot path was not.\\n\\n## Changes\\n\\n- **`mcp_config_custom.go`** — collapsed both branches into a single call to `ReplaceTemplateExpressionsWithEnvVars`, which produces `\\\\${VAR}` for all JSON engines. Bash strips the leading `\\\\` in the heredoc, leaving a literal `${VAR}` that the gateway resolves from its own environment after the JSON is parsed. Removed the now-unused `requiresCopilotFields` param from `renderCustomMCPEnvVars`. Updated doc comment with the heredoc injection rationale.\\n\\n- **`mcp_config_copilot_test.go`** — added `TestRenderCustomMCPEnvVars_NonCopilotSecretsEscaped` covering non-Copilot secrets, Copilot secrets, and fallback expressions (`${{ secrets.X || 'default' }}`), asserting `\\\\${VAR}` and asserting the absence of `${VAR}`.\\n\\n## Before / after\\n\\n```\\n# Before (non-Copilot): bash expands ${PROMETHEUS_PASSWORD} → raw value spliced into JSON\\n\\\"env\\\": {\\n  \\\"PROMETHEUS_PASSWORD\\\": \\\"${PROMETHEUS_PASSWORD}\\\"   ← unquoted heredoc expands this\\n}\\n\\n# After: literal \\\\${PROMETHEUS_PASSWORD} survives heredoc intact as valid JSON\\n\\\"env\\\": {\\n  \\\"PROMETHEUS_PASSWORD\\\": \\\"\\\\${PROMETHEUS_PASSWORD}\\\"  ← gateway substitutes safely\\n}\\n```\\n\\nThe `\\\\${VAR}` pattern was already used correctly for HTTP header secrets and the built-in safe-outputs MCP server in the same generated file; this aligns the custom env block path with the rest.\\n\\n<!-- START COPILOT CODING AGENT SUFFIX -->\\n\\n- Fixes #49424\",\"changedFiles\":2,\"deletions\":12,\"files\":[{\"path\":\"pkg/workflow/mcp_config_copilot_test.go\",\"additions\":111,\"deletions\":0,\"changeType\":\"MODIFIED\"},{\"path\":\"pkg/workflow/mcp_config_custom.go\",\"additions\":16,\"deletions\":12,\"changeType\":\"MODIFIED\"}],\"headRefName\":\"copilot/fix-stdio-container-env-secrets\",\"headRefOid\":\"b9ca441354ffefa90a01abec8bb31656ba995fb5\",\"number\":49450,\"title\":\"fix: escape MCP env secrets with \\\\${VAR} for non-Copilot engines to prevent heredoc JSON corruption\"}\n---\n176 /tmp/gh-aw/agent/pr-diff.patch\n---\n[\n  {\n    \"body\": \"This removes the only production call to `ReplaceSecretsWithBashVars`, but that exported helper remains documented as being used for non-Copilot MCP env blocks (`secret_extraction.go:165-168`). Please\",\n    \"id\": 3694593524,\n    \"line\": 81,\n    \"path\": \"pkg/workflow/mcp_config_custom.go\",\n    \"user\": \"Copilot\"\n  }\n]\n---cache---\nspdd-daily\n<shellId: 11 completed with exit code 0>","detailedContent":"{\"additions\":127,\"body\":\"For non-Copilot engines (Claude, Codex, etc.), secrets in stdio/container custom MCP server `env:` blocks were substituted as bare `${VAR}` inside an unquoted heredoc. Bash expands those before the MCP gateway parses the config, so any secret containing `\\\"` or `\\\\` silently corrupts the JSON and the **Start MCP Gateway** step fails.\\n\\n## Root cause\\n\\n`renderCustomMCPEnvVars` had separate branches for Copilot (`\\\\${VAR}`, heredoc-safe) and non-Copilot (`${VAR}`, bash-expanded). The Copilot path was correct; the non-Copilot path was not.\\n\\n## Changes\\n\\n- **`mcp_config_custom.go`** — collapsed both branches into a single call to `ReplaceTemplateExpressionsWithEnvVars`, which produces `\\\\${VAR}` for all JSON engines. Bash strips the leading `\\\\` in the heredoc, leaving a literal `${VAR}` that the gateway resolves from its own environment after the JSON is parsed. Removed the now-unused `requiresCopilotFields` param from `renderCustomMCPEnvVars`. Updated doc comment with the heredoc injection rationale.\\n\\n- **`mcp_config_copilot_test.go`** — added `TestRenderCustomMCPEnvVars_NonCopilotSecretsEscaped` covering non-Copilot secrets, Copilot secrets, and fallback expressions (`${{ secrets.X || 'default' }}`), asserting `\\\\${VAR}` and asserting the absence of `${VAR}`.\\n\\n## Before / after\\n\\n```\\n# Before (non-Copilot): bash expands ${PROMETHEUS_PASSWORD} → raw value spliced into JSON\\n\\\"env\\\": {\\n  \\\"PROMETHEUS_PASSWORD\\\": \\\"${PROMETHEUS_PASSWORD}\\\"   ← unquoted heredoc expands this\\n}\\n\\n# After: literal \\\\${PROMETHEUS_PASSWORD} survives heredoc intact as valid JSON\\n\\\"env\\\": {\\n  \\\"PROMETHEUS_PASSWORD\\\": \\\"\\\\${PROMETHEUS_PASSWORD}\\\"  ← gateway substitutes safely\\n}\\n```\\n\\nThe `\\\\${VAR}` pattern was already used correctly for HTTP header secrets and the built-in safe-outputs MCP server in the same generated file; this aligns the custom env block path with the rest.\\n\\n<!-- START COPILOT CODING AGENT SUFFIX -->\\n\\n- Fixes #49424\",\"changedFiles\":2,\"deletions\":12,\"files\":[{\"path\":\"pkg/workflow/mcp_config_copilot_test.go\",\"additions\":111,\"deletions\":0,\"changeType\":\"MODIFIED\"},{\"path\":\"pkg/workflow/mcp_config_custom.go\",\"additions\":16,\"deletions\":12,\"changeType\":\"MODIFIED\"}],\"headRefName\":\"copilot/fix-stdio-container-env-secrets\",\"headRefOid\":\"b9ca441354ffefa90a01abec8bb31656ba995fb5\",\"number\":49450,\"title\":\"fix: escape MCP env secrets with \\\\${VAR} for non-Copilot engines to prevent heredoc JSON corruption\"}\n---\n176 /tmp/gh-aw/agent/pr-diff.patch\n---\n[\n  {\n    \"body\": \"This removes the only production call to `ReplaceSecretsWithBashVars`, but that exported helper remains documented as being used for non-Copilot MCP env blocks (`secret_extraction.go:165-168`). Please\",\n    \"id\": 3694593524,\n    \"line\": 81,\n    \"path\": \"pkg/workflow/mcp_config_custom.go\",\n    \"user\": \"Copilot\"\n  }\n]\n---cache---\nspdd-daily\n<shellId: 11 completed with exit code 0>"}}}
{"type":"assistant.message","timestamp":"2026-08-01T06:25:55.528Z","data":{"content":""}}
{"type":"tool.execution_start","timestamp":"2026-08-01T06:25:55.531Z","data":{"toolName":"view","mcpServerName":""}}
{"type":"tool.execution_complete","timestamp":"2026-08-01T06:25:55.585Z","data":{"toolName":"view","mcpServerName":"","success":true,"result":{"content":"1. diff --git a/pkg/workflow/mcp_config_copilot_test.go b/pkg/workflow/mcp_config_copilot_test.go\n2. index 5d953df9320..22c78c9b09c 100644\n3. --- a/pkg/workflow/mcp_config_copilot_test.go\n4. +++ b/pkg/workflow/mcp_config_copilot_test.go\n5. @@ -203,6 +203,117 @@ func TestRenderSharedMCPConfig_ToolsFieldGeneration(t *testing.T) {\n6.  \t}\n7.  }\n8.  \n9. +// TestRenderCustomMCPEnvVars_NonCopilotSecretsEscaped verifies that for non-Copilot\n10. +// JSON engines, secrets in custom MCP server env blocks are rendered as \\${VAR}\n11. +// (backslash-escaped) rather than ${VAR} (unescaped). Unescaped references would\n12. +// be expanded by bash inside the unquoted heredoc that carries the MCP gateway\n13. +// JSON config -- a secret containing '\"' or '\\' would corrupt the JSON and cause\n14. +// the gateway to fail before the agent runs. Backslash-escaping keeps the JSON\n15. +// valid regardless of the secret's runtime value.\n16. +func TestRenderCustomMCPEnvVars_NonCopilotSecretsEscaped(t *testing.T) {\n17. +\ttests := []struct {\n18. +\t\tname              string\n19. +\t\ttoolConfig        map[string]any\n20. +\t\trenderer          MCPConfigRenderer\n21. +\t\texpectedContent   []string\n22. +\t\tunexpectedContent []string\n23. +\t}{\n24. +\t\t{\n25. +\t\t\tname: \"Non-Copilot stdio container - secret in env uses backslash-escaped var\",\n26. +\t\t\ttoolConfig: map[string]any{\n27. +\t\t\t\t\"type\":      \"stdio\",\n28. +\t\t\t\t\"container\": \"some/image:latest\",\n29. +\t\t\t\t\"env\": map[string]any{\n30. +\t\t\t\t\t\"MY_TOKEN\": \"${{ secrets.MY_TOKEN }}\",\n31. +\t\t\t\t},\n32. +\t\t\t},\n33. +\t\t\trenderer: MCPConfigRenderer{\n34. +\t\t\t\tIndentLevel:           \"  \",\n35. +\t\t\t\tFormat:                \"json\",\n36. +\t\t\t\tRequiresCopilotFields: false,\n37. +\t\t\t},\n38. +\t\t\t// Secret must be rendered as \\${MY_TOKEN} so the unquoted heredoc\n39. +\t\t\t// leaves a literal ${MY_TOKEN} string in the JSON (valid JSON).\n40. +\t\t\texpectedContent: []string{\n41. +\t\t\t\t`\"MY_TOKEN\": \"\\${MY_TOKEN}\"`,\n42. +\t\t\t},\n43. +\t\t\t// Must NOT appear as an unescaped bash variable reference -- that\n44. +\t\t\t// would let bash splice the raw secret value into the JSON.\n45. +\t\t\tunexpectedContent: []string{\n46. +\t\t\t\t`\"MY_TOKEN\": \"${MY_TOKEN}\"`,\n47. +\t\t\t},\n48. +\t\t},\n49. +\t\t{\n50. +\t\t\tname: \"Copilot stdio - secret in env also uses backslash-escaped var\",\n51. +\t\t\ttoolConfig: map[string]any{\n52. +\t\t\t\t\"type\":      \"stdio\",\n53. +\t\t\t\t\"container\": \"some/image:latest\",\n54. +\t\t\t\t\"env\": map[string]any{\n55. +\t\t\t\t\t\"MY_TOKEN\": \"${{ secrets.MY_TOKEN }}\",\n56. +\t\t\t\t},\n57. +\t\t\t\t\"allowed\": []string{\"*\"},\n58. +\t\t\t},\n59. +\t\t\trenderer: MCPConfigRenderer{\n60. +\t\t\t\tIndentLevel:           \"  \",\n61. +\t\t\t\tFormat:                \"json\",\n62. +\t\t\t\tRequiresCopilotFields: true,\n63. +\t\t\t},\n64. +\t\t\texpectedContent: []string{\n65. +\t\t\t\t`\"MY_TOKEN\": \"\\${MY_TOKEN}\"`,\n66. +\t\t\t},\n67. +\t\t\tunexpectedContent: []string{\n68. +\t\t\t\t`\"MY_TOKEN\": \"${MY_TOKEN}\"`,\n69. +\t\t\t},\n70. +\t\t},\n71. +\t\t{\n72. +\t\t\tname: \"Non-Copilot stdio container - secret with fallback in env uses backslash-escaped var\",\n73. +\t\t\ttoolConfig: map[string]any{\n74. +\t\t\t\t\"type\":      \"stdio\",\n75. +\t\t\t\t\"container\": \"some/image:latest\",\n76. +\t\t\t\t\"env\": map[string]any{\n77. +\t\t\t\t\t\"DD_SITE\": \"${{ secrets.DD_SITE || 'datadoghq.com' }}\",\n78. +\t\t\t\t},\n79. +\t\t\t},\n80. +\t\t\trenderer: MCPConfigRenderer{\n81. +\t\t\t\tIndentLevel:           \"  \",\n82. +\t\t\t\tFormat:                \"json\",\n83. +\t\t\t\tRequiresCopilotFields: false,\n84. +\t\t\t},\n85. +\t\t\texpectedContent: []string{\n86. +\t\t\t\t`\"DD_SITE\": \"\\${DD_SITE}\"`,\n87. +\t\t\t},\n88. +\t\t\tunexpectedContent: []string{\n89. +\t\t\t\t`\"DD_SITE\": \"${DD_SITE}\"`,\n90. +\t\t\t},\n91. +\t\t},\n92. +\t}\n93. +\n94. +\tfor _, tt := range tests {\n95. +\t\tt.Run(tt.name, func(t *testing.T) {\n96. +\t\t\tvar output strings.Builder\n97. +\n98. +\t\t\terr := renderSharedMCPConfig(&output, \"test-tool\", tt.toolConfig, tt.renderer)\n99. +\t\t\tif err != nil {\n100. +\t\t\t\tt.Fatalf(\"renderSharedMCPConfig failed: %v\", err)\n101. +\t\t\t}\n102. +\n103. +\t\t\tresult := output.String()\n104. +\n105. +\t\t\tfor _, expected := range tt.expectedContent {\n106. +\t\t\t\tif !strings.Contains(result, expected) {\n107. +\t\t\t\t\tt.Errorf(\"Expected content not found: %q\\nActual output:\\n%s\", expected, result)\n108. +\t\t\t\t}\n109. +\t\t\t}\n110. +\n111. +\t\t\tfor _, unexpected := range tt.unexpectedContent {\n112. +\t\t\t\tif strings.Contains(result, unexpected) {\n113. +\t\t\t\t\tt.Errorf(\"Unexpected content found: %q\\nActual output:\\n%s\", unexpected, result)\n114. +\t\t\t\t}\n115. +\t\t\t}\n116. +\t\t})\n117. +\t}\n118. +}\n119. +\n120.  func TestRenderSharedMCPConfig_TypeConversion(t *testing.T) {\n121.  \ttests := []struct {\n122.  \t\tname           string\n123. diff --git a/pkg/workflow/mcp_config_custom.go b/pkg/workflow/mcp_config_custom.go\n124. index b4f2d05c575..ccbe1b92230 100644\n125. --- a/pkg/workflow/mcp_config_custom.go\n126. +++ b/pkg/workflow/mcp_config_custom.go\n127. @@ -54,10 +54,16 @@ func renderCustomMCPConfigWrapperWithContext(yaml *strings.Builder, toolName str\n128.  //\n129.  // For TOML output, GitHub Actions template expressions are rewritten to direct\n130.  // ${VAR} references because Codex config expects shell-style environment\n131. -// expansion. For JSON output, Copilot uses escaped \\${VAR} passthrough syntax,\n132. -// while non-Copilot engines use bash variable substitution to avoid embedding\n133. -// secret expressions directly in the generated run block.\n134. -func renderCustomMCPEnvVars(env map[string]string, tomlFormat, requiresCopilotFields bool) map[string]string {\n135. +// expansion. For JSON output, both Copilot and non-Copilot engines use the\n136. +// escaped \\${VAR} passthrough syntax. The MCP gateway JSON config is written\n137. +// inside an unquoted heredoc, so any unescaped ${VAR} reference would be\n138. +// expanded by bash before the gateway sees it — splicing the raw secret bytes\n139. +// into JSON text. A secret value containing a '\"' or '\\' character would\n140. +// corrupt the JSON. Backslash-escaping (\\${VAR}) prevents the heredoc from\n141. +// expanding the variable; bash only strips the leading backslash, leaving the\n142. +// literal ${VAR} string in the JSON, which the gateway then resolves safely\n143. +// from its own environment (RGS-008 compliance).\n144. +func renderCustomMCPEnvVars(env map[string]string, tomlFormat bool) map[string]string {\n145.  \trenderedEnv := make(map[string]string, len(env))\n146.  \tfor envKey, envValue := range env {\n147.  \t\tif tomlFormat {\n148. @@ -67,14 +73,12 @@ func renderCustomMCPEnvVars(env map[string]string, tomlFormat, requiresCopilotFi\n149.  \t\t\tenvValue = strings.ReplaceAll(envValue, \"${{ env.\", \"${\")\n150.  \t\t\tenvValue = strings.ReplaceAll(envValue, \"${{ github.workspace }}\", \"${GITHUB_WORKSPACE}\")\n151.  \t\t\tenvValue = strings.ReplaceAll(envValue, \" }}\", \"}\")\n152. -\t\t} else if requiresCopilotFields {\n153. -\t\t\t// For Copilot, replace all template expressions with \\${VAR} syntax.\n154. -\t\t\tenvValue = ReplaceTemplateExpressionsWithEnvVars(envValue)\n155.  \t\t} else {\n156. -\t\t\t// For non-Copilot engines, replace secrets with ${VAR} bash expansion so\n157. -\t\t\t// they are never directly interpolated in the run block (RGS-008). The\n158. -\t\t\t// env vars are injected into the step env block by collectMCPEnvironmentVariables.\n159. -\t\t\tenvValue = ReplaceSecretsWithBashVars(envValue)\n160. +\t\t\t// For both Copilot and non-Copilot JSON engines, replace all template\n161. +\t\t\t// expressions with \\${VAR} passthrough syntax. This keeps raw secret values\n162. +\t\t\t// out of the heredoc (RGS-008) and produces valid JSON regardless of the\n163. +\t\t\t// characters contained in the secret at runtime.\n164. +\t\t\tenvValue = ReplaceTemplateExpressionsWithEnvVars(envValue)\n165.  \t\t}\n166.  \t\trenderedEnv[envKey] = envValue\n167.  \t}\n168. @@ -338,7 +342,7 @@ func renderMCPMapProperty(yaml *strings.Builder, property string, isLast bool, m\n169.  }\n170.  \n171.  func renderMCPEnvMap(yaml *strings.Builder, isLast bool, mcpConfig *parser.RegistryMCPServerConfig, renderer MCPConfigRenderer, headerSecrets map[string]string) {\n172. -\trenderedEnv := renderCustomMCPEnvVars(mcpConfig.Env, renderer.Format == \"toml\", renderer.RequiresCopilotFields)\n173. +\trenderedEnv := renderCustomMCPEnvVars(mcpConfig.Env, renderer.Format == \"toml\")\n174.  \tif renderer.Format == \"toml\" {\n175.  \t\twriteTOMLInlineStringMapSection(yaml, renderer.IndentLevel, \"env\", renderedEnv)\n176.  \t\treturn\n177. ","detailedContent":"\ndiff --git a/tmp/gh-aw/agent/pr-diff.patch b/tmp/gh-aw/agent/pr-diff.patch\nindex 0000000..0000000 100644\n--- a/tmp/gh-aw/agent/pr-diff.patch\n+++ b/tmp/gh-aw/agent/pr-diff.patch\n@@ -1,177 +1,177 @@\n diff --git a/pkg/workflow/mcp_config_copilot_test.go b/pkg/workflow/mcp_config_copilot_test.go\n index 5d953df9320..22c78c9b09c 100644\n --- a/pkg/workflow/mcp_config_copilot_test.go\n +++ b/pkg/workflow/mcp_config_copilot_test.go\n @@ -203,6 +203,117 @@ func TestRenderSharedMCPConfig_ToolsFieldGeneration(t *testing.T) {\n  \t}\n  }\n  \n +// TestRenderCustomMCPEnvVars_NonCopilotSecretsEscaped verifies that for non-Copilot\n +// JSON engines, secrets in custom MCP server env blocks are rendered as \\${VAR}\n +// (backslash-escaped) rather than ${VAR} (unescaped). Unescaped references would\n +// be expanded by bash inside the unquoted heredoc that carries the MCP gateway\n +// JSON config -- a secret containing '\"' or '\\' would corrupt the JSON and cause\n +// the gateway to fail before the agent runs. Backslash-escaping keeps the JSON\n +// valid regardless of the secret's runtime value.\n +func TestRenderCustomMCPEnvVars_NonCopilotSecretsEscaped(t *testing.T) {\n +\ttests := []struct {\n +\t\tname              string\n +\t\ttoolConfig        map[string]any\n +\t\trenderer          MCPConfigRenderer\n +\t\texpectedContent   []string\n +\t\tunexpectedContent []string\n +\t}{\n +\t\t{\n +\t\t\tname: \"Non-Copilot stdio container - secret in env uses backslash-escaped var\",\n +\t\t\ttoolConfig: map[string]any{\n +\t\t\t\t\"type\":      \"stdio\",\n +\t\t\t\t\"container\": \"some/image:latest\",\n +\t\t\t\t\"env\": map[string]any{\n +\t\t\t\t\t\"MY_TOKEN\": \"${{ secrets.MY_TOKEN }}\",\n +\t\t\t\t},\n +\t\t\t},\n +\t\t\trenderer: MCPConfigRenderer{\n +\t\t\t\tIndentLevel:           \"  \",\n +\t\t\t\tFormat:                \"json\",\n +\t\t\t\tRequiresCopilotFields: false,\n +\t\t\t},\n +\t\t\t// Secret must be rendered as \\${MY_TOKEN} so the unquoted heredoc\n +\t\t\t// leaves a literal ${MY_TOKEN} string in the JSON (valid JSON).\n +\t\t\texpectedContent: []string{\n +\t\t\t\t`\"MY_TOKEN\": \"\\${MY_TOKEN}\"`,\n +\t\t\t},\n +\t\t\t// Must NOT appear as an unescaped bash variable reference -- that\n +\t\t\t// would let bash splice the raw secret value into the JSON.\n +\t\t\tunexpectedContent: []string{\n +\t\t\t\t`\"MY_TOKEN\": \"${MY_TOKEN}\"`,\n +\t\t\t},\n +\t\t},\n +\t\t{\n +\t\t\tname: \"Copilot stdio - secret in env also uses backslash-escaped var\",\n +\t\t\ttoolConfig: map[string]any{\n +\t\t\t\t\"type\":      \"stdio\",\n +\t\t\t\t\"container\": \"some/image:latest\",\n +\t\t\t\t\"env\": map[string]any{\n +\t\t\t\t\t\"MY_TOKEN\": \"${{ secrets.MY_TOKEN }}\",\n +\t\t\t\t},\n +\t\t\t\t\"allowed\": []string{\"*\"},\n +\t\t\t},\n +\t\t\trenderer: MCPConfigRenderer{\n +\t\t\t\tIndentLevel:           \"  \",\n +\t\t\t\tFormat:                \"json\",\n +\t\t\t\tRequiresCopilotFields: true,\n +\t\t\t},\n +\t\t\texpectedContent: []string{\n +\t\t\t\t`\"MY_TOKEN\": \"\\${MY_TOKEN}\"`,\n +\t\t\t},\n +\t\t\tunexpectedContent: []string{\n +\t\t\t\t`\"MY_TOKEN\": \"${MY_TOKEN}\"`,\n +\t\t\t},\n +\t\t},\n +\t\t{\n +\t\t\tname: \"Non-Copilot stdio container - secret with fallback in env uses backslash-escaped var\",\n +\t\t\ttoolConfig: map[string]any{\n +\t\t\t\t\"type\":      \"stdio\",\n +\t\t\t\t\"container\": \"some/image:latest\",\n +\t\t\t\t\"env\": map[string]any{\n +\t\t\t\t\t\"DD_SITE\": \"${{ secrets.DD_SITE || 'datadoghq.com' }}\",\n +\t\t\t\t},\n +\t\t\t},\n +\t\t\trenderer: MCPConfigRenderer{\n +\t\t\t\tIndentLevel:           \"  \",\n +\t\t\t\tFormat:                \"json\",\n +\t\t\t\tRequiresCopilotFields: false,\n +\t\t\t},\n +\t\t\texpectedContent: []string{\n +\t\t\t\t`\"DD_SITE\": \"\\${DD_SITE}\"`,\n +\t\t\t},\n +\t\t\tunexpectedContent: []string{\n +\t\t\t\t`\"DD_SITE\": \"${DD_SITE}\"`,\n +\t\t\t},\n +\t\t},\n +\t}\n +\n +\tfor _, tt := range tests {\n +\t\tt.Run(tt.name, func(t *testing.T) {\n +\t\t\tvar output strings.Builder\n +\n +\t\t\terr := renderSharedMCPConfig(&output, \"test-tool\", tt.toolConfig, tt.renderer)\n +\t\t\tif err != nil {\n +\t\t\t\tt.Fatalf(\"renderSharedMCPConfig failed: %v\", err)\n +\t\t\t}\n +\n +\t\t\tresult := output.String()\n +\n +\t\t\tfor _, expected := range tt.expectedContent {\n +\t\t\t\tif !strings.Contains(result, expected) {\n +\t\t\t\t\tt.Errorf(\"Expected content not found: %q\\nActual output:\\n%s\", expected, result)\n +\t\t\t\t}\n +\t\t\t}\n +\n +\t\t\tfor _, unexpected := range tt.unexpectedContent {\n +\t\t\t\tif strings.Contains(result, unexpected) {\n +\t\t\t\t\tt.Errorf(\"Unexpected content found: %q\\nActual output:\\n%s\", unexpected, result)\n +\t\t\t\t}\n +\t\t\t}\n +\t\t})\n +\t}\n +}\n +\n  func TestRenderSharedMCPConfig_TypeConversion(t *testing.T) {\n  \ttests := []struct {\n  \t\tname           string\n diff --git a/pkg/workflow/mcp_config_custom.go b/pkg/workflow/mcp_config_custom.go\n index b4f2d05c575..ccbe1b92230 100644\n --- a/pkg/workflow/mcp_config_custom.go\n +++ b/pkg/workflow/mcp_config_custom.go\n @@ -54,10 +54,16 @@ func renderCustomMCPConfigWrapperWithContext(yaml *strings.Builder, toolName str\n  //\n  // For TOML output, GitHub Actions template expressions are rewritten to direct\n  // ${VAR} references because Codex config expects shell-style environment\n -// expansion. For JSON output, Copilot uses escaped \\${VAR} passthrough syntax,\n -// while non-Copilot engines use bash variable substitution to avoid embedding\n -// secret expressions directly in the generated run block.\n -func renderCustomMCPEnvVars(env map[string]string, tomlFormat, requiresCopilotFields bool) map[string]string {\n +// expansion. For JSON output, both Copilot and non-Copilot engines use the\n +// escaped \\${VAR} passthrough syntax. The MCP gateway JSON config is written\n +// inside an unquoted heredoc, so any unescaped ${VAR} reference would be\n +// expanded by bash before the gateway sees it — splicing the raw secret bytes\n +// into JSON text. A secret value containing a '\"' or '\\' character would\n +// corrupt the JSON. Backslash-escaping (\\${VAR}) prevents the heredoc from\n +// expanding the variable; bash only strips the leading backslash, leaving the\n +// literal ${VAR} string in the JSON, which the gateway then resolves safely\n +// from its own environment (RGS-008 compliance).\n +func renderCustomMCPEnvVars(env map[string]string, tomlFormat bool) map[string]string {\n  \trenderedEnv := make(map[string]string, len(env))\n  \tfor envKey, envValue := range env {\n  \t\tif tomlFormat {\n @@ -67,14 +73,12 @@ func renderCustomMCPEnvVars(env map[string]string, tomlFormat, requiresCopilotFi\n  \t\t\tenvValue = strings.ReplaceAll(envValue, \"${{ env.\", \"${\")\n  \t\t\tenvValue = strings.ReplaceAll(envValue, \"${{ github.workspace }}\", \"${GITHUB_WORKSPACE}\")\n  \t\t\tenvValue = strings.ReplaceAll(envValue, \" }}\", \"}\")\n -\t\t} else if requiresCopilotFields {\n -\t\t\t// For Copilot, replace all template expressions with \\${VAR} syntax.\n -\t\t\tenvValue = ReplaceTemplateExpressionsWithEnvVars(envValue)\n  \t\t} else {\n -\t\t\t// For non-Copilot engines, replace secrets with ${VAR} bash expansion so\n -\t\t\t// they are never directly interpolated in the run block (RGS-008). The\n -\t\t\t// env vars are injected into the step env block by collectMCPEnvironmentVariables.\n -\t\t\tenvValue = ReplaceSecretsWithBashVars(envValue)\n +\t\t\t// For both Copilot and non-Copilot JSON engines, replace all template\n +\t\t\t// expressions with \\${VAR} passthrough syntax. This keeps raw secret values\n +\t\t\t// out of the heredoc (RGS-008) and produces valid JSON regardless of the\n +\t\t\t// characters contained in the secret at runtime.\n +\t\t\tenvValue = ReplaceTemplateExpressionsWithEnvVars(envValue)\n  \t\t}\n  \t\trenderedEnv[envKey] = envValue\n  \t}\n @@ -338,7 +342,7 @@ func renderMCPMapProperty(yaml *strings.Builder, property string, isLast bool, m\n  }\n  \n  func renderMCPEnvMap(yaml *strings.Builder, isLast bool, mcpConfig *parser.RegistryMCPServerConfig, renderer MCPConfigRenderer, headerSecrets map[string]string) {\n -\trenderedEnv := renderCustomMCPEnvVars(mcpConfig.Env, renderer.Format == \"toml\", renderer.RequiresCopilotFields)\n +\trenderedEnv := renderCustomMCPEnvVars(mcpConfig.Env, renderer.Format == \"toml\")\n  \tif renderer.Format == \"toml\" {\n  \t\twriteTOMLInlineStringMapSection(yaml, renderer.IndentLevel, \"env\", renderedEnv)\n  \t\treturn\n \n"}}}
{"type":"assistant.message","timestamp":"2026-08-01T06:26:12.953Z","data":{"content":""}}
{"type":"tool.execution_start","timestamp":"2026-08-01T06:26:12.957Z","data":{"toolName":"task","mcpServerName":""}}
{"type":"subagent.started","timestamp":"2026-08-01T06:26:12.965Z","data":{"agentName":"general-purpose","agentDisplayName":"General Purpose Agent","toolCallId":"toolu_01Rx3W5kmkRUPFVAAxvCugZe"}}
{"type":"tool.execution_complete","timestamp":"2026-08-01T06:26:12.971Z","data":{"toolName":"task","mcpServerName":"","success":true,"result":{"content":"Agent started in background with agent_id: grumpy-coder-3. You'll be notified when it completes. Tell the user you're waiting and end your response, or continue unrelated work until notified. The agent supports multi-turn conversations — use write_agent to send follow-up messages.","detailedContent":"Prompt to general-purpose agent (grumpy-coder-3):\nYou are a highly critical, grumpy senior Go reviewer doing FIRST-PASS issue mining on a pull request diff in the repo github/gh-aw (Go codebase at /home/runner/work/gh-aw/gh-aw).\n\nPR #49450: \"fix: escape MCP env secrets with \\${VAR} for non-Copilot engines to prevent heredoc JSON corruption\"\n\nContext: `renderCustomMCPEnvVars` in pkg/workflow/mcp_config_custom.go previously had two branches: Copilot engines used `ReplaceTemplateExpressionsWithEnvVars` (producing `\\${VAR}`), non-Copilot engines used `ReplaceSecretsWithBashVars` (producing `${VAR}`, expanded by bash in an unquoted heredoc — a bug because secrets containing `\"` or `\\` corrupt JSON). The fix collapses both branches to always use `ReplaceTemplateExpressionsWithEnvVars`, and removes the now-unused `requiresCopilotFields` parameter from `renderCustomMCPEnvVars`.\n\nThe diff (176 lines) is at /tmp/gh-aw/agent/pr-diff.patch — read it directly. It touches:\n1. pkg/workflow/mcp_config_copilot_test.go — new test TestRenderCustomMCPEnvVars_NonCopilotSecretsEscaped\n2. pkg/workflow/mcp_config_custom.go — renderCustomMCPEnvVars signature change and body change\n\nYour job:\n1. Read /tmp/gh-aw/agent/pr-diff.patch.\n2. Read the full current content of pkg/workflow/mcp_config_custom.go and pkg/workflow/mcp_config_copilot_test.go in the repo to get full context around the changed lines (not just the diff hunk).\n3. Also check pkg/workflow/secret_extraction.go — specifically the ReplaceSecretsWithBashVars and ReplaceTemplateExpressionsWithEnvVars functions and their doc comments — to see if ReplaceSecretsWithBashVars is now completely unused in production code (only used in tests), and if its doc comment is now stale/misleading given this PR removes its only production call site.\n4. Search the whole repo for other callers of ReplaceSecretsWithBashVars and requiresCopilotFields-like patterns to make sure no other place still relies on the removed behavior (e.g. collectMCPEnvironmentVariables which the removed code comment mentioned \"env vars are injected into the step env block by collectMCPEnvironmentVariables\" — verify this still holds true for the new unified code path, i.e., check that the env var injection mechanism still functions correctly for non-Copilot engines now that they use \\${VAR} instead of ${VAR}).\n5. Look for logic errors, edge cases, dead code, stale comments, weak test coverage in the new test (e.g., does it cover TOML/Codex path, does it cover the `RequiresCopilotFields` field even though it no longer affects behavior — is that field now entirely vestigial/dead and should be removed elsewhere, check its other usages), and any regressions in behavior for Codex/TOML engines.\n\nOutput STRICT JSONL — one JSON object per line, each with fields:\n{\"severity\": \"critical|high|medium|low\", \"file\": \"path/relative/to/repo\", \"line\": <line number in the file, best estimate>, \"issue\": \"short description\", \"impact\": \"why it matters\"}\n\nOnly report issues within the actual changed lines/functions of this diff (mcp_config_custom.go's renderCustomMCPEnvVars/renderMCPEnvMap, and the new test file), or directly caused by this change (e.g., stale docs on ReplaceSecretsWithBashVars now that it's unused in production, or a now-vestigial RequiresCopilotFields field). Do not comment on unrelated pre-existing code. Do not include markdown fences or any prose outside the JSONL lines. If no valid issues found, output nothing."}}}
[copilot-sdk-driver] [sdk-driver] error: Execution failed: Error: No model available. Check policy enablement under GitHub Settings > Copilot
{"type":"subagent.completed","timestamp":"2026-08-01T06:26:13.026Z","data":{"agentName":"general-purpose","toolCallId":"toolu_01Rx3W5kmkRUPFVAAxvCugZe"}}

Action Required

Assign this issue to an agent to debug and fix the issue.

Debug with any coding agent

Use this prompt with any coding agent (GitHub Copilot, Claude, Gemini, etc.):

Debug the agentic workflow failure using https://raw.githubusercontent.com/github/gh-aw/main/debug.md

The failed workflow run is at https://github.com/github/gh-aw/actions/runs/30683628677

Manually invoke the agent

Debug this workflow failure using your favorite Agent CLI and the agentic-workflows prompt.

• Start your agent
• Load the agentic-workflows skill from .github/skills/agentic-workflows/SKILL.md or https://github.com/github/gh-aw/blob/main/.github/skills/agentic-workflows/SKILL.md
• Type debug the agentic workflow pr-code-quality-reviewer failure in https://github.com/github/gh-aw/actions/runs/30683628677

Tip

Stop reporting this workflow as a failure

To stop a workflow from creating failure issues, set report-failure-as-issue: false in its frontmatter:

safe-outputs:
  report-failure-as-issue: false

Generated from PR Code Quality Reviewer · 55.4 AIC · ◷

• expires on Aug 1, 2026, 6:30 PM UTC

[github-actions]

Agent job 30687802684 failed.

Warning

Engine Failure: The copilot engine terminated unexpectedly.

Last agent output:

{"type":"tool.execution_start","timestamp":"2026-08-01T06:59:48.078Z","data":{"toolName":"bash","mcpServerName":"","command":"printf '{\"path\":\"pkg/workflow/add_labels.go\",\"line\":36,\"body\":\"No validation prevents `issues: false` together with `pull-requests: false`, silently minting **zero** permissions for add-labels.\",\"start_line\":33,\"side\":\"RIGHT\"}' | safeoutputs create_pull_request_review_comment . 2>&1 | tail -30"}}
{"type":"tool.execution_complete","timestamp":"2026-08-01T06:59:48.515Z","data":{"toolName":"bash","mcpServerName":"","success":true,"result":{"content":"[info] [safeoutputs] Bridge invoked: url=(awmgmcpg/redacted) toolsFile=/home/runner/work/_temp/gh-aw/mcp-cli/tools/safeoutputs.json, userArgs=[\"create_pull_request_review_comment\",\".\"]\n[info] [safeoutputs] Calling tool 'create_pull_request_review_comment' with args: {\"path\":\"pkg/workflow/add_labels.go\",\"line\":36,\"body\":\"No validation prevents `issues: false` together with `pull-requests: false`, silently minting **zero** permissions for add-labels.\",\"start_line\":33,\"side\":\"RIGHT\"}\n[info] [safeoutputs] MCP initialize: POST (awmgmcpg/redacted) [safeoutputs] MCP initialize: status=200, sessionId=WVB3YWHI..., elapsed=23ms\n[info] [safeoutputs] MCP notifications/initialized\n[info] [safeoutputs] MCP notifications/initialized: done (6ms)\n[info] [safeoutputs] MCP keepalive started (interval=10000ms)\n[info] [safeoutputs] MCP tools/call: tool=create_pull_request_review_comment, args={\"path\":\"pkg/workflow/add_labels.go\",\"line\":36,\"body\":\"No validation prevents `issues: false` together with `pull-requests: false`, silently minting **zero** permissions for add-labels.\",\"start_line\":33,\"side\":\"RIGHT\"}\n[info] [safeoutputs] MCP tools/call: status=200, elapsed=275ms\n[info] [safeoutputs] MCP keepalive stopped\n[info] [safeoutputs] Tool call complete: total=305ms\n{\"result\":\"success\"}\n[info] [safeoutputs] Tool output: 20 chars\n<shellId: 77 completed with exit code 0>","detailedContent":"[info] [safeoutputs] Bridge invoked: url=(awmgmcpg/redacted) toolsFile=/home/runner/work/_temp/gh-aw/mcp-cli/tools/safeoutputs.json, userArgs=[\"create_pull_request_review_comment\",\".\"]\n[info] [safeoutputs] Calling tool 'create_pull_request_review_comment' with args: {\"path\":\"pkg/workflow/add_labels.go\",\"line\":36,\"body\":\"No validation prevents `issues: false` together with `pull-requests: false`, silently minting **zero** permissions for add-labels.\",\"start_line\":33,\"side\":\"RIGHT\"}\n[info] [safeoutputs] MCP initialize: POST (awmgmcpg/redacted) [safeoutputs] MCP initialize: status=200, sessionId=WVB3YWHI..., elapsed=23ms\n[info] [safeoutputs] MCP notifications/initialized\n[info] [safeoutputs] MCP notifications/initialized: done (6ms)\n[info] [safeoutputs] MCP keepalive started (interval=10000ms)\n[info] [safeoutputs] MCP tools/call: tool=create_pull_request_review_comment, args={\"path\":\"pkg/workflow/add_labels.go\",\"line\":36,\"body\":\"No validation prevents `issues: false` together with `pull-requests: false`, silently minting **zero** permissions for add-labels.\",\"start_line\":33,\"side\":\"RIGHT\"}\n[info] [safeoutputs] MCP tools/call: status=200, elapsed=275ms\n[info] [safeoutputs] MCP keepalive stopped\n[info] [safeoutputs] Tool call complete: total=305ms\n{\"result\":\"success\"}\n[info] [safeoutputs] Tool output: 20 chars\n<shellId: 77 completed with exit code 0>"}}}
{"type":"assistant.message","timestamp":"2026-08-01T06:59:55.780Z","data":{"content":"Good, that worked. Now the second key comment on the target/pull-requests cross-validation gap."}}
{"type":"tool.execution_start","timestamp":"2026-08-01T06:59:55.782Z","data":{"toolName":"bash","mcpServerName":"","command":"printf '%s' '{\"path\":\"pkg/workflow/add_labels.go\",\"line\":36,\"start_line\":33,\"side\":\"RIGHT\",\"body\":\"Disabling `pull-requests` has no compile-time link to `target`; if the workflow can label PRs (e.g. `target: \\\"*\\\"` or triggered on a PR) but sets `pull-requests: false`, the runtime `issues.addLabels` call will fail with a permissions error the compiler never warns about.\\n\\n<details><summary>💡 Details</summary>\\n\\nGitHub'\"'\"'s REST docs list `POST /repos/{owner}/{repo}/issues/{issue_number}/labels` as requiring both `issues:write` **and** `pull_requests:write` for GitHub App tokens when the target could be a PR (a PR is an issue under the hood). This PR lets users independently opt out of either permission without validating that against `target`/trigger context, so `add-labels` + `pull-requests: false` + a PR-eligible target is a real foot-gun that will only surface at runtime, not at `gh aw compile` time.\\n\\nSuggested mitigation: add a validation check (e.g., in `safe_outputs_validation.go`) that errors/warns when `pull-requests: false` is combined with a `target` value that can resolve to a pull request, similar to how the analogous `add-comment` config shares this same gap today.\\n\\n</details>\"}' | safeoutputs create_pull_request_review_comment ."}}
{"type":"tool.execution_complete","timestamp":"2026-08-01T06:59:56.017Z","data":{"toolName":"bash","mcpServerName":"","success":true,"result":{"content":"[info] [safeoutputs] Bridge invoked: url=(awmgmcpg/redacted) toolsFile=/home/runner/work/_temp/gh-aw/mcp-cli/tools/safeoutputs.json, userArgs=[\"create_pull_request_review_comment\",\".\"]\n[info] [safeoutputs] Calling tool 'create_pull_request_review_comment' with args: {\"path\":\"pkg/workflow/add_labels.go\",\"line\":36,\"start_line\":33,\"side\":\"RIGHT\",\"body\":\"Disabling `pull-requests` has no compile-time link to `target`; if the workflow can label PRs (e.g. `target: \\\"*\\\"` or triggered on a PR) but sets `pull-requests: false`, the runtime `issues.addLabels` call will fail with a permissions error the compiler never warns about.\\n\\n<details><summary>💡 Details</summary>\\n\\nGitHub's REST docs list `POST /repos/{owner}/{repo}/issues/{issue_number}/labels` as requiring both `issues:write` **and** `pull_requests:write` for GitHub App tokens when the target could be a PR (a PR is an issue under the hood). This PR lets users independently opt out of either permission without validating that against `target`/trigger context, so `add-labels` + `pull-requests: false` + a PR-eligible target is a real foot-gun that will only surface at runtime, not at `gh aw compile` time.\\n\\nSuggested mitigation: add a validation check (e.g., in `safe_outputs_validation.go`) that errors/warns when `pull-requests: false` is combined with a `target` value that can resolve to a pull request, similar to how the analogous `add-comment` config shares this same gap today.\\n\\n</details>\"}\n[info] [safeoutputs] MCP initialize: POST (awmgmcpg/redacted) [safeoutputs] MCP initialize: status=200, sessionId=MT3ADOJR..., elapsed=24ms\n[info] [safeoutputs] MCP notifications/initialized\n[info] [safeoutputs] MCP notifications/initialized: done (6ms)\n[info] [safeoutputs] MCP keepalive started (interval=10000ms)\n[info] [safeoutputs] MCP tools/call: tool=create_pull_request_review_comment, args={\"path\":\"pkg/workflow/add_labels.go\",\"line\":36,\"start_line\":33,\"side\":\"RIGHT\",\"body\":\"Disabling `pull-requests` has no compile-time link to `target`; if the workflow can label PRs (e.g. `target: \\\"*\\\"` or triggered on a PR) but sets `pull-requests: false`, the runtime `issues.addLabels` call will fail with a permissions error the compiler never warns about.\\n\\n<details><summary>💡 Details</summary>\\n\\nGitHub's REST docs list `POST /repos/{owner}/{repo}/issues/{issue_number}/labels` as requiring both `issues:write` **and** `pull_requests:write` for GitHub App tokens when the target could be a PR (a PR is an issue under the hood). This PR lets users independently opt out of either permission without validating that against `target`/trigger context, so `add-labels` + `pull-requests: false` + a PR-eligible target is a real foot-gun that will only surface at runtime, not at `gh aw compile` time.\\n\\nSuggested mitigation: add a validation check (e.g., in `safe_outputs_validation.go`) that errors/warns when `pull-requests: false` is combined with a `target` value that can resolve to a pull request, similar to how the analogous `add-comment` config shares this same gap today.\\n\\n</details>\"}\n[info] [safeoutputs] MCP tools/call: status=200, elapsed=31ms\n[info] [safeoutputs] MCP keepalive stopped\n[info] [safeoutputs] Tool call complete: total=62ms\n{\"result\":\"success\"}\n[info] [safeoutputs] Tool output: 20 chars\n<shellId: 78 completed with exit code 0>","detailedContent":"[info] [safeoutputs] Bridge invoked: url=(awmgmcpg/redacted) toolsFile=/home/runner/work/_temp/gh-aw/mcp-cli/tools/safeoutputs.json, userArgs=[\"create_pull_request_review_comment\",\".\"]\n[info] [safeoutputs] Calling tool 'create_pull_request_review_comment' with args: {\"path\":\"pkg/workflow/add_labels.go\",\"line\":36,\"start_line\":33,\"side\":\"RIGHT\",\"body\":\"Disabling `pull-requests` has no compile-time link to `target`; if the workflow can label PRs (e.g. `target: \\\"*\\\"` or triggered on a PR) but sets `pull-requests: false`, the runtime `issues.addLabels` call will fail with a permissions error the compiler never warns about.\\n\\n<details><summary>💡 Details</summary>\\n\\nGitHub's REST docs list `POST /repos/{owner}/{repo}/issues/{issue_number}/labels` as requiring both `issues:write` **and** `pull_requests:write` for GitHub App tokens when the target could be a PR (a PR is an issue under the hood). This PR lets users independently opt out of either permission without validating that against `target`/trigger context, so `add-labels` + `pull-requests: false` + a PR-eligible target is a real foot-gun that will only surface at runtime, not at `gh aw compile` time.\\n\\nSuggested mitigation: add a validation check (e.g., in `safe_outputs_validation.go`) that errors/warns when `pull-requests: false` is combined with a `target` value that can resolve to a pull request, similar to how the analogous `add-comment` config shares this same gap today.\\n\\n</details>\"}\n[info] [safeoutputs] MCP initialize: POST (awmgmcpg/redacted) [safeoutputs] MCP initialize: status=200, sessionId=MT3ADOJR..., elapsed=24ms\n[info] [safeoutputs] MCP notifications/initialized\n[info] [safeoutputs] MCP notifications/initialized: done (6ms)\n[info] [safeoutputs] MCP keepalive started (interval=10000ms)\n[info] [safeoutputs] MCP tools/call: tool=create_pull_request_review_comment, args={\"path\":\"pkg/workflow/add_labels.go\",\"line\":36,\"start_line\":33,\"side\":\"RIGHT\",\"body\":\"Disabling `pull-requests` has no compile-time link to `target`; if the workflow can label PRs (e.g. `target: \\\"*\\\"` or triggered on a PR) but sets `pull-requests: false`, the runtime `issues.addLabels` call will fail with a permissions error the compiler never warns about.\\n\\n<details><summary>💡 Details</summary>\\n\\nGitHub's REST docs list `POST /repos/{owner}/{repo}/issues/{issue_number}/labels` as requiring both `issues:write` **and** `pull_requests:write` for GitHub App tokens when the target could be a PR (a PR is an issue under the hood). This PR lets users independently opt out of either permission without validating that against `target`/trigger context, so `add-labels` + `pull-requests: false` + a PR-eligible target is a real foot-gun that will only surface at runtime, not at `gh aw compile` time.\\n\\nSuggested mitigation: add a validation check (e.g., in `safe_outputs_validation.go`) that errors/warns when `pull-requests: false` is combined with a `target` value that can resolve to a pull request, similar to how the analogous `add-comment` config shares this same gap today.\\n\\n</details>\"}\n[info] [safeoutputs] MCP tools/call: status=200, elapsed=31ms\n[info] [safeoutputs] MCP keepalive stopped\n[info] [safeoutputs] Tool call complete: total=62ms\n{\"result\":\"success\"}\n[info] [safeoutputs] Tool output: 20 chars\n<shellId: 78 completed with exit code 0>"}}}
{"type":"assistant.message","timestamp":"2026-08-01T07:00:05.108Z","data":{"content":"Both key findings posted. Now submitting the overall review."}}
{"type":"tool.execution_start","timestamp":"2026-08-01T07:00:05.111Z","data":{"toolName":"bash","mcpServerName":"","command":"printf '%s' '{\"event\":\"COMMENT\",\"body\":\"### Review: least-privilege opt-out for add-labels permissions\\n\\nApproach mirrors the existing `add-comment` permission-toggle pattern and is implemented consistently, but it inherits (and slightly extends) a validation gap: nothing ties the new `issues`/`pull-requests` opt-outs to `target`, so `pull-requests: false` combined with a PR-eligible target can compile cleanly and fail at runtime with a GitHub API permissions error. Also, setting both flags false silently yields a zero-permission handler with no warning.\\n\\n<details><summary>💡 Themes</summary>\\n\\n- **Structural consistency**: `buildAddLabelsPermissions` closely mirrors `buildAddCommentPermissions` (added in a very recent prior commit), so this change follows established convention rather than introducing a new pattern.\\n- **Runtime vs compile-time gap**: GitHub'\"'\"'s REST API for adding labels requires both `issues:write` and `pull-requests:write` when the target could be a PR. No cross-validation exists between these new opt-out flags and the `target` field, so a misconfigured workflow will only fail when it actually runs, not when it is compiled.\\n- **Both-false edge case**: setting `issues: false` and `pull-requests: false` together produces a technically valid but functionally useless (zero-permission) handler with no compiler feedback.\\n- **Test coverage**: new tests cover each flag in isolation (`pull-requests: false` alone, `issues: false` alone) plus an app-token end-to-end case, but do not cover the combined-false or target-interaction scenarios called out above.\\n\\nNot blocking merge since this mirrors an already-accepted pattern in `add-comment`, but flagged as a real correctness gap worth addressing (ideally for both call sites) in a follow-up.\\n\\n</details>\"}' | safeoutputs submit_pull_request_review ."}}
{"type":"tool.execution_complete","timestamp":"2026-08-01T07:00:05.358Z","data":{"toolName":"bash","mcpServerName":"","success":true,"result":{"content":"[info] [safeoutputs] Bridge invoked: url=(awmgmcpg/redacted) toolsFile=/home/runner/work/_temp/gh-aw/mcp-cli/tools/safeoutputs.json, userArgs=[\"submit_pull_request_review\",\".\"]\n[info] [safeoutputs] Calling tool 'submit_pull_request_review' with args: {\"event\":\"COMMENT\",\"body\":\"### Review: least-privilege opt-out for add-labels permissions\\n\\nApproach mirrors the existing `add-comment` permission-toggle pattern and is implemented consistently, but it inherits (and slightly extends) a validation gap: nothing ties the new `issues`/`pull-requests` opt-outs to `target`, so `pull-requests: false` combined with a PR-eligible target can compile cleanly and fail at runtime with a GitHub API permissions error. Also, setting both flags false silently yields a zero-permission handler with no warning.\\n\\n<details><summary>💡 Themes</summary>\\n\\n- **Structural consistency**: `buildAddLabelsPermissions` closely mirrors `buildAddCommentPermissions` (added in a very recent prior commit), so this change follows established convention rather than introducing a new pattern.\\n- **Runtime vs compile-time gap**: GitHub's REST API for adding labels requires both `issues:write` and `pull-requests:write` when the target could be a PR. No cross-validation exists between these new opt-out flags and the `target` field, so a misconfigured workflow will only fail when it actually runs, not when it is compiled.\\n- **Both-false edge case**: setting `issues: false` and `pull-requests: false` together produces a technically valid but functionally useless (zero-permission) handler with no compiler feedback.\\n- **Test coverage**: new tests cover each flag in isolation (`pull-requests: false` alone, `issues: false` alone) plus an app-token end-to-end case, but do not cover the combined-false or target-interaction scenarios called out above.\\n\\nNot blocking merge since this mirrors an already-accepted pattern in `add-comment`, but flagged as a real correctness gap worth addressing (ideally for both call sites) in a follow-up.\\n\\n</details>\"}\n[info] [safeoutputs] MCP initialize: POST (awmgmcpg/redacted) [safeoutputs] MCP initialize: status=200, sessionId=LPHD72CS..., elapsed=24ms\n[info] [safeoutputs] MCP notifications/initialized\n[info] [safeoutputs] MCP notifications/initialized: done (10ms)\n[info] [safeoutputs] MCP keepalive started (interval=10000ms)\n[info] [safeoutputs] MCP tools/call: tool=submit_pull_request_review, args={\"event\":\"COMMENT\",\"body\":\"### Review: least-privilege opt-out for add-labels permissions\\n\\nApproach mirrors the existing `add-comment` permission-toggle pattern and is implemented consistently, but it inherits (and slightly extends) a validation gap: nothing ties the new `issues`/`pull-requests` opt-outs to `target`, so `pull-requests: false` combined with a PR-eligible target can compile cleanly and fail at runtime with a GitHub API permissions error. Also, setting both flags false silently yields a zero-permission handler with no warning.\\n\\n<details><summary>💡 Themes</summary>\\n\\n- **Structural consistency**: `buildAddLabelsPermissions` closely mirrors `buildAddCommentPermissions` (added in a very recent prior commit), so this change follows established convention rather than introducing a new pattern.\\n- **Runtime vs compile-time gap**: GitHub's REST API for adding labels requires both `issues:write` and `pull-requests:write` when the target could be a PR. No cross-validation exists between these new opt-out flags and the `target` field, so a misconfigured workflow will only fail when it actually runs, not when it is compiled.\\n- **Both-false edge case**: setting `issues: false` and `pull-requests: false` together produces a technically valid but functionally useless (zero-permission) handler with no compiler feedback.\\n- **Test coverage**: new tests cover each flag in isolation (`pull-requests: false` alone, `issues: false` alone) plus an app-token end-to-end case, but do not cover the combined-false or target-interaction scenarios called out above.\\n\\nNot blocking merge since this mirrors an already-accepted pattern in `add-comment`, but flagged as a real correctness gap worth addressing (ideally for both call sites) in a follow-up.\\n\\n</details>\"}\n[info] [safeoutputs] MCP tools/call: status=200, elapsed=31ms\n[info] [safeoutputs] MCP keepalive stopped\n[info] [safeoutputs] Tool call complete: total=67ms\n{\"result\":\"success\"}\n[info] [safeoutputs] Tool output: 20 chars\n<shellId: 79 completed with exit code 0>","detailedContent":"[info] [safeoutputs] Bridge invoked: url=(awmgmcpg/redacted) toolsFile=/home/runner/work/_temp/gh-aw/mcp-cli/tools/safeoutputs.json, userArgs=[\"submit_pull_request_review\",\".\"]\n[info] [safeoutputs] Calling tool 'submit_pull_request_review' with args: {\"event\":\"COMMENT\",\"body\":\"### Review: least-privilege opt-out for add-labels permissions\\n\\nApproach mirrors the existing `add-comment` permission-toggle pattern and is implemented consistently, but it inherits (and slightly extends) a validation gap: nothing ties the new `issues`/`pull-requests` opt-outs to `target`, so `pull-requests: false` combined with a PR-eligible target can compile cleanly and fail at runtime with a GitHub API permissions error. Also, setting both flags false silently yields a zero-permission handler with no warning.\\n\\n<details><summary>💡 Themes</summary>\\n\\n- **Structural consistency**: `buildAddLabelsPermissions` closely mirrors `buildAddCommentPermissions` (added in a very recent prior commit), so this change follows established convention rather than introducing a new pattern.\\n- **Runtime vs compile-time gap**: GitHub's REST API for adding labels requires both `issues:write` and `pull-requests:write` when the target could be a PR. No cross-validation exists between these new opt-out flags and the `target` field, so a misconfigured workflow will only fail when it actually runs, not when it is compiled.\\n- **Both-false edge case**: setting `issues: false` and `pull-requests: false` together produces a technically valid but functionally useless (zero-permission) handler with no compiler feedback.\\n- **Test coverage**: new tests cover each flag in isolation (`pull-requests: false` alone, `issues: false` alone) plus an app-token end-to-end case, but do not cover the combined-false or target-interaction scenarios called out above.\\n\\nNot blocking merge since this mirrors an already-accepted pattern in `add-comment`, but flagged as a real correctness gap worth addressing (ideally for both call sites) in a follow-up.\\n\\n</details>\"}\n[info] [safeoutputs] MCP initialize: POST (awmgmcpg/redacted) [safeoutputs] MCP initialize: status=200, sessionId=LPHD72CS..., elapsed=24ms\n[info] [safeoutputs] MCP notifications/initialized\n[info] [safeoutputs] MCP notifications/initialized: done (10ms)\n[info] [safeoutputs] MCP keepalive started (interval=10000ms)\n[info] [safeoutputs] MCP tools/call: tool=submit_pull_request_review, args={\"event\":\"COMMENT\",\"body\":\"### Review: least-privilege opt-out for add-labels permissions\\n\\nApproach mirrors the existing `add-comment` permission-toggle pattern and is implemented consistently, but it inherits (and slightly extends) a validation gap: nothing ties the new `issues`/`pull-requests` opt-outs to `target`, so `pull-requests: false` combined with a PR-eligible target can compile cleanly and fail at runtime with a GitHub API permissions error. Also, setting both flags false silently yields a zero-permission handler with no warning.\\n\\n<details><summary>💡 Themes</summary>\\n\\n- **Structural consistency**: `buildAddLabelsPermissions` closely mirrors `buildAddCommentPermissions` (added in a very recent prior commit), so this change follows established convention rather than introducing a new pattern.\\n- **Runtime vs compile-time gap**: GitHub's REST API for adding labels requires both `issues:write` and `pull-requests:write` when the target could be a PR. No cross-validation exists between these new opt-out flags and the `target` field, so a misconfigured workflow will only fail when it actually runs, not when it is compiled.\\n- **Both-false edge case**: setting `issues: false` and `pull-requests: false` together produces a technically valid but functionally useless (zero-permission) handler with no compiler feedback.\\n- **Test coverage**: new tests cover each flag in isolation (`pull-requests: false` alone, `issues: false` alone) plus an app-token end-to-end case, but do not cover the combined-false or target-interaction scenarios called out above.\\n\\nNot blocking merge since this mirrors an already-accepted pattern in `add-comment`, but flagged as a real correctness gap worth addressing (ideally for both call sites) in a follow-up.\\n\\n</details>\"}\n[info] [safeoutputs] MCP tools/call: status=200, elapsed=31ms\n[info] [safeoutputs] MCP keepalive stopped\n[info] [safeoutputs] Tool call complete: total=67ms\n{\"result\":\"success\"}\n[info] [safeoutputs] Tool output: 20 chars\n<shellId: 79 completed with exit code 0>"}}}
{"type":"assistant.message","timestamp":"2026-08-01T07:00:09.108Z","data":{"content":"Review complete. Posted 2 review comments on `pkg/workflow/add_labels.go` flagging: (1) missing validation for both `issues`/`pull-requests` set false → silent zero-permission handler, and (2) no cross-validation between `pull-requests: false` and PR-eligible `target` values, which GitHub's API requires together — a real runtime-failure risk the compiler won't catch. Submitted overall review as `COMMENT` (non-blocking, since it mirrors an already-accepted pattern in `add-comment`) with a themes summary in `<details>`."}}
[copilot-sdk-driver] [sdk-driver] session completed: hasOutput=true assistantTurns=4 durationMs=844875

Generated from PR Code Quality Reviewer · 193.3 AIC · ◷

[github-actions]

Agent job 30689558152 failed.

Warning

Engine Failure: The copilot engine terminated unexpectedly.

Last agent output:

{"type":"tool.execution_complete","timestamp":"2026-08-01T07:47:47.532Z","data":{"toolName":"bash","mcpServerName":"","success":true,"result":{"content":"},\n  },\n  defaultOptions: [],\n  create(context) {\n    const sourceCode = context.sourceCode;\n    type Scope = ReturnType<typeof sourceCode.getScope>;\n\n    function isDefinitionAvailableAtNode(definition: TSESLint.Scope.Definition, node: TSESTree.Node): boolean {\n      if (definition.type === \"ImportBinding\" || definition.type === \"FunctionName\") {\n        return true;\n      }\n      const definitionNode = definition.name ?? definition.node;\n      // range is absent on synthetic/virtual nodes; treat as unavailable rather than throwing\n      if (!definitionNode?.range || !node.range) return false;\n      return definitionNode.range[0] < node.range[0];\n    }\n\n    function hasResolvableLocalBinding(node: TSESTree.Node, name: string): boolean {\n      let scope: Scope | null = sourceCode.getScope(node);\n      while (scope) {\n        const variable = scope.set.get(name);\n        if (variable && variable.defs.some(def => isDefinitionAvailableAtNode(def, node))) {\n          return true;\n        }\n        scope = scope.upper;\n      }\n      return false;\n    }\n\n    return {\n      TemplateLiteral(node) {\n        // Tagged templates pass values to the tag function as-is; they are not\n        // string-coerced by interpolation, so flagging them would be incorrect.\n        if (node.parent?.type === AST_NODE_TYPES.TaggedTemplateExpression) return;\n\n        for (const expr of node.expressions) {\n          if (!isBareIdentifierExpression(expr)) continue;\n\n          // Resolve the identifier through the full scope chain. This correctly\n          // handles closures: an identifier inside a nested function can still\n          // resolve to an outer catch binding.\n          let scope: Scope | null = sourceCode.getScope(expr);\n          let variable: TSESLint.Scope.Variable | null = null;\n          while (scope) {\n            const v = scope.set.get(expr.name);\n            if (v) {\n              variable = v;\n              break;\n            }\n            scope = scope.upper;\n          }\n<shellId: 14 completed with exit code 0>","detailedContent":"},\n  },\n  defaultOptions: [],\n  create(context) {\n    const sourceCode = context.sourceCode;\n    type Scope = ReturnType<typeof sourceCode.getScope>;\n\n    function isDefinitionAvailableAtNode(definition: TSESLint.Scope.Definition, node: TSESTree.Node): boolean {\n      if (definition.type === \"ImportBinding\" || definition.type === \"FunctionName\") {\n        return true;\n      }\n      const definitionNode = definition.name ?? definition.node;\n      // range is absent on synthetic/virtual nodes; treat as unavailable rather than throwing\n      if (!definitionNode?.range || !node.range) return false;\n      return definitionNode.range[0] < node.range[0];\n    }\n\n    function hasResolvableLocalBinding(node: TSESTree.Node, name: string): boolean {\n      let scope: Scope | null = sourceCode.getScope(node);\n      while (scope) {\n        const variable = scope.set.get(name);\n        if (variable && variable.defs.some(def => isDefinitionAvailableAtNode(def, node))) {\n          return true;\n        }\n        scope = scope.upper;\n      }\n      return false;\n    }\n\n    return {\n      TemplateLiteral(node) {\n        // Tagged templates pass values to the tag function as-is; they are not\n        // string-coerced by interpolation, so flagging them would be incorrect.\n        if (node.parent?.type === AST_NODE_TYPES.TaggedTemplateExpression) return;\n\n        for (const expr of node.expressions) {\n          if (!isBareIdentifierExpression(expr)) continue;\n\n          // Resolve the identifier through the full scope chain. This correctly\n          // handles closures: an identifier inside a nested function can still\n          // resolve to an outer catch binding.\n          let scope: Scope | null = sourceCode.getScope(expr);\n          let variable: TSESLint.Scope.Variable | null = null;\n          while (scope) {\n            const v = scope.set.get(expr.name);\n            if (v) {\n              variable = v;\n              break;\n            }\n            scope = scope.upper;\n          }\n<shellId: 14 completed with exit code 0>"}}}
{"type":"assistant.message","timestamp":"2026-08-01T07:47:51.207Z","data":{"content":"This confirms the code is a byte-for-byte replication of the existing pattern used in three sibling rules (identical `isDefinitionAvailableAtNode`/`hasResolvableLocalBinding` implementation). The pre-existing Copilot bot comment about TDZ ordering applies equally to all four rules — it's a pre-existing shared issue, not something introduced/worsened by this diff, so it's not a new blocking regression in this PR."}}
{"type":"tool.execution_start","timestamp":"2026-08-01T07:47:51.209Z","data":{"toolName":"bash","mcpServerName":"","command":"cd /home/runner/work/gh-aw/gh-aw && sed -n '1,120p' eslint-factory/src/rules/no-err-stack-then-string-fallback.ts"}}
{"type":"tool.execution_complete","timestamp":"2026-08-01T07:47:51.279Z","data":{"toolName":"bash","mcpServerName":"","success":true,"result":{"content":"import { AST_NODE_TYPES, ESLintUtils, TSESLint, TSESTree } from \"`@typescript-eslint/utils`\";\n\nconst createRule = ESLintUtils.RuleCreator(name => `https://github.com/github/gh-aw/tree/main/eslint-factory#${name}`);\n\nfunction isIdentifierNamed(node: TSESTree.Node, name: string): node is TSESTree.Identifier {\n  return node.type === AST_NODE_TYPES.Identifier && node.name === name;\n}\n\n/**\n * Checks whether `node` matches `<errVar> && <errVar>.stack` (LogicalExpression).\n */\nfunction isErrAndErrStack(node: TSESTree.Node, errVar: string): boolean {\n  if (node.type !== AST_NODE_TYPES.LogicalExpression || node.operator !== \"&&\") return false;\n  if (!isIdentifierNamed(node.left, errVar)) return false;\n  const right = node.right;\n  if (right.type !== AST_NODE_TYPES.MemberExpression || right.computed) return false;\n  return isIdentifierNamed(right.object, errVar) && isIdentifierNamed(right.property, \"stack\");\n}\n\n/**\n * Checks whether `node` matches `<errVar> instanceof Error` (BinaryExpression).\n */\nfunction isErrInstanceofError(node: TSESTree.Node, errVar: string): boolean {\n  if (node.type !== AST_NODE_TYPES.BinaryExpression || node.operator !== \"instanceof\") return false;\n  return isIdentifierNamed(node.left, errVar) && isIdentifierNamed(node.right, \"Error\");\n}\n\n/**\n * Checks whether `node` matches `<errVar>.stack` (MemberExpression).\n */\nfunction isErrStack(node: TSESTree.Node, errVar: string): boolean {\n  if (node.type !== AST_NODE_TYPES.MemberExpression || node.computed) return false;\n  return isIdentifierNamed(node.object, errVar) && isIdentifierNamed(node.property, \"stack\");\n}\n\n/**\n * Checks whether `node` matches `String(<errVar>)`.\n */\nfunction isStringErr(node: TSESTree.Node, errVar: string): boolean {\n  if (node.type !== AST_NODE_TYPES.CallExpression) return false;\n  if (!isIdentifierNamed(node.callee, \"String\")) return false;\n  if (node.arguments.length !== 1) return false;\n  return isIdentifierNamed(node.arguments[0], errVar);\n}\n\nfunction isDefinitionAvailableAtNode(definition: TSESLint.Scope.Definition, node: TSESTree.Node): boolean {\n  if (definition.type === \"ImportBinding\" || definition.type === \"FunctionName\") {\n    return true;\n  }\n  const definitionNode = definition.name ?? definition.node;\n  if (!definitionNode?.range || !node.range) return false;\n  return definitionNode.range[0] < node.range[0];\n}\n\nexport const noErrStackThenStringFallbackRule = createRule({\n  name: \"no-err-stack-then-string-fallback\",\n  meta: {\n    type: \"suggestion\",\n    hasSuggestions: true,\n    docs: {\n      description:\n        \"Prefer getErrorMessage(err) over `err && err.stack ? err.stack : String(err)` or \" +\n        \"`err instanceof Error ? err.stack : String(err)`. \" +\n        \"The stack-trace form surfaces noisy implementation details; getErrorMessage() returns \" +\n        \"the concise error message and is available in every actions/setup/js script via error_helpers.cjs.\",\n    },\n    schema: [],\n    messages: {\n      preferGetErrorMessage: \"Prefer getErrorMessage({{errorVar}}) from error_helpers.cjs. The `{{errorVar}}.stack` ternary surfaces noisy stack frames; getErrorMessage() returns a clean, consistent message.\",\n      replaceWithGetErrorMessage: \"Replace with getErrorMessage({{errorVar}}) — ensure getErrorMessage is imported from error_helpers.cjs before applying.\",\n    },\n  },\n  defaultOptions: [],\n  create(context) {\n    const sourceCode = context.sourceCode;\n    type SourceCodeScope = ReturnType<typeof sourceCode.getScope>;\n\n    function hasResolvableLocalBinding(node: TSESTree.Node, name: string): boolean {\n      let scope: SourceCodeScope | null = sourceCode.getScope(node);\n      while (scope) {\n        const variable = scope.set.get(name);\n        if (variable && variable.defs.some(def => isDefinitionAvailableAtNode(def, node))) {\n          return true;\n        }\n        scope = scope.upper;\n      }\n      return false;\n    }\n\n    return {\n      ConditionalExpression(node) {\n        // Patterns:\n        //   <errVar> && <errVar>.stack ? <errVar>.stack : String(<errVar>)\n        //   <errVar> instanceof Error  ? <errVar>.stack : String(<errVar>)\n        const test = node.test;\n\n        let errVar: string;\n        if (test.type === AST_NODE_TYPES.LogicalExpression) {\n          // Resolve errVar from the test's left-hand side identifier\n          if (test.left.type !== AST_NODE_TYPES.Identifier) return;\n          errVar = test.left.name;\n          if (!isErrAndErrStack(test, errVar)) return;\n        } else if (test.type === AST_NODE_TYPES.BinaryExpression) {\n          if (test.left.type !== AST_NODE_TYPES.Identifier) return;\n          errVar = test.left.name;\n          if (!isErrInstanceofError(test, errVar)) return;\n        } else {\n          return;\n        }\n\n        if (!isErrStack(node.consequent, errVar)) return;\n        if (!isStringErr(node.alternate, errVar)) return;\n\n        context.report({\n          node,\n          messageId: \"preferGetErrorMessage\",\n          data: { errorVar: errVar },\n          suggest: hasResolvableLocalBinding(node, \"getErrorMessage\")\n            ? [\n                {\n<shellId: 15 completed with exit code 0>","detailedContent":"import { AST_NODE_TYPES, ESLintUtils, TSESLint, TSESTree } from \"`@typescript-eslint/utils`\";\n\nconst createRule = ESLintUtils.RuleCreator(name => `https://github.com/github/gh-aw/tree/main/eslint-factory#${name}`);\n\nfunction isIdentifierNamed(node: TSESTree.Node, name: string): node is TSESTree.Identifier {\n  return node.type === AST_NODE_TYPES.Identifier && node.name === name;\n}\n\n/**\n * Checks whether `node` matches `<errVar> && <errVar>.stack` (LogicalExpression).\n */\nfunction isErrAndErrStack(node: TSESTree.Node, errVar: string): boolean {\n  if (node.type !== AST_NODE_TYPES.LogicalExpression || node.operator !== \"&&\") return false;\n  if (!isIdentifierNamed(node.left, errVar)) return false;\n  const right = node.right;\n  if (right.type !== AST_NODE_TYPES.MemberExpression || right.computed) return false;\n  return isIdentifierNamed(right.object, errVar) && isIdentifierNamed(right.property, \"stack\");\n}\n\n/**\n * Checks whether `node` matches `<errVar> instanceof Error` (BinaryExpression).\n */\nfunction isErrInstanceofError(node: TSESTree.Node, errVar: string): boolean {\n  if (node.type !== AST_NODE_TYPES.BinaryExpression || node.operator !== \"instanceof\") return false;\n  return isIdentifierNamed(node.left, errVar) && isIdentifierNamed(node.right, \"Error\");\n}\n\n/**\n * Checks whether `node` matches `<errVar>.stack` (MemberExpression).\n */\nfunction isErrStack(node: TSESTree.Node, errVar: string): boolean {\n  if (node.type !== AST_NODE_TYPES.MemberExpression || node.computed) return false;\n  return isIdentifierNamed(node.object, errVar) && isIdentifierNamed(node.property, \"stack\");\n}\n\n/**\n * Checks whether `node` matches `String(<errVar>)`.\n */\nfunction isStringErr(node: TSESTree.Node, errVar: string): boolean {\n  if (node.type !== AST_NODE_TYPES.CallExpression) return false;\n  if (!isIdentifierNamed(node.callee, \"String\")) return false;\n  if (node.arguments.length !== 1) return false;\n  return isIdentifierNamed(node.arguments[0], errVar);\n}\n\nfunction isDefinitionAvailableAtNode(definition: TSESLint.Scope.Definition, node: TSESTree.Node): boolean {\n  if (definition.type === \"ImportBinding\" || definition.type === \"FunctionName\") {\n    return true;\n  }\n  const definitionNode = definition.name ?? definition.node;\n  if (!definitionNode?.range || !node.range) return false;\n  return definitionNode.range[0] < node.range[0];\n}\n\nexport const noErrStackThenStringFallbackRule = createRule({\n  name: \"no-err-stack-then-string-fallback\",\n  meta: {\n    type: \"suggestion\",\n    hasSuggestions: true,\n    docs: {\n      description:\n        \"Prefer getErrorMessage(err) over `err && err.stack ? err.stack : String(err)` or \" +\n        \"`err instanceof Error ? err.stack : String(err)`. \" +\n        \"The stack-trace form surfaces noisy implementation details; getErrorMessage() returns \" +\n        \"the concise error message and is available in every actions/setup/js script via error_helpers.cjs.\",\n    },\n    schema: [],\n    messages: {\n      preferGetErrorMessage: \"Prefer getErrorMessage({{errorVar}}) from error_helpers.cjs. The `{{errorVar}}.stack` ternary surfaces noisy stack frames; getErrorMessage() returns a clean, consistent message.\",\n      replaceWithGetErrorMessage: \"Replace with getErrorMessage({{errorVar}}) — ensure getErrorMessage is imported from error_helpers.cjs before applying.\",\n    },\n  },\n  defaultOptions: [],\n  create(context) {\n    const sourceCode = context.sourceCode;\n    type SourceCodeScope = ReturnType<typeof sourceCode.getScope>;\n\n    function hasResolvableLocalBinding(node: TSESTree.Node, name: string): boolean {\n      let scope: SourceCodeScope | null = sourceCode.getScope(node);\n      while (scope) {\n        const variable = scope.set.get(name);\n        if (variable && variable.defs.some(def => isDefinitionAvailableAtNode(def, node))) {\n          return true;\n        }\n        scope = scope.upper;\n      }\n      return false;\n    }\n\n    return {\n      ConditionalExpression(node) {\n        // Patterns:\n        //   <errVar> && <errVar>.stack ? <errVar>.stack : String(<errVar>)\n        //   <errVar> instanceof Error  ? <errVar>.stack : String(<errVar>)\n        const test = node.test;\n\n        let errVar: string;\n        if (test.type === AST_NODE_TYPES.LogicalExpression) {\n          // Resolve errVar from the test's left-hand side identifier\n          if (test.left.type !== AST_NODE_TYPES.Identifier) return;\n          errVar = test.left.name;\n          if (!isErrAndErrStack(test, errVar)) return;\n        } else if (test.type === AST_NODE_TYPES.BinaryExpression) {\n          if (test.left.type !== AST_NODE_TYPES.Identifier) return;\n          errVar = test.left.name;\n          if (!isErrInstanceofError(test, errVar)) return;\n        } else {\n          return;\n        }\n\n        if (!isErrStack(node.consequent, errVar)) return;\n        if (!isStringErr(node.alternate, errVar)) return;\n\n        context.report({\n          node,\n          messageId: \"preferGetErrorMessage\",\n          data: { errorVar: errVar },\n          suggest: hasResolvableLocalBinding(node, \"getErrorMessage\")\n            ? [\n                {\n<shellId: 15 completed with exit code 0>"}}}
{"type":"assistant.message","timestamp":"2026-08-01T07:48:07.330Z","data":{"content":"Now let me start the grumpy-coder sub-agent while I finish my own analysis, and check test file completeness."}}
{"type":"tool.execution_start","timestamp":"2026-08-01T07:48:07.332Z","data":{"toolName":"task","mcpServerName":""}}
{"type":"subagent.started","timestamp":"2026-08-01T07:48:07.340Z","data":{"agentName":"general-purpose","agentDisplayName":"General Purpose Agent","toolCallId":"toolu_01DU3A9boP1zkM7asH8wxwqi"}}
{"type":"tool.execution_complete","timestamp":"2026-08-01T07:48:07.346Z","data":{"toolName":"task","mcpServerName":"","success":true,"result":{"content":"Agent started in background with agent_id: grumpy-coder-3. You'll be notified when it completes. Tell the user you're waiting and end your response, or continue unrelated work until notified. The agent supports multi-turn conversations — use write_agent to send follow-up messages.","detailedContent":"Prompt to general-purpose agent (grumpy-coder-3):\nYou are a highly critical, grumpy senior code reviewer. Review this PR diff (github/gh-aw PR #49501) for correctness risks, performance regressions, maintainability debt, and weak engineering. Only consider CHANGED lines in the diff below.\n\nPR summary: gates the `replaceWithGetErrorMessage` autofix suggestion in the `no-err-stack-then-string-fallback` ESLint rule on whether `getErrorMessage` is resolvable in local scope (identical pattern already used in 3 sibling rules: no-json-stringify-error, prefer-get-error-message-over-string, no-caught-error-interpolation).\n\nDiff:\n```\ndiff --git a/eslint-factory/src/rules/no-err-stack-then-string-fallback.ts b/eslint-factory/src/rules/no-err-stack-then-string-fallback.ts\nindex 0af67d2118d..ef526b1687b 100644\n--- a/eslint-factory/src/rules/no-err-stack-then-string-fallback.ts\n+++ b/eslint-factory/src/rules/no-err-stack-then-string-fallback.ts\n@@ -1,4 +1,4 @@\n-import { AST_NODE_TYPES, ESLintUtils, TSESTree } from \"`@typescript-eslint/utils`\";\n+import { AST_NODE_TYPES, ESLintUtils, TSESLint, TSESTree } from \"`@typescript-eslint/utils`\";\n \n const createRule = ESLintUtils.RuleCreator(name => `https://github.com/github/gh-aw/tree/main/eslint-factory#${name}`);\n \n@@ -43,6 +43,15 @@ function isStringErr(node: TSESTree.Node, errVar: string): boolean {\n   return isIdentifierNamed(node.arguments[0], errVar);\n }\n \n+function isDefinitionAvailableAtNode(definition: TSESLint.Scope.Definition, node: TSESTree.Node): boolean {\n+  if (definition.type === \"ImportBinding\" || definition.type === \"FunctionName\") {\n+    return true;\n+  }\n+  const definitionNode = definition.name ?? definition.node;\n+  if (!definitionNode?.range || !node.range) return false;\n+  return definitionNode.range[0] < node.range[0];\n+}\n+\n export const noErrStackThenStringFallbackRule = createRule({\n   name: \"no-err-stack-then-string-fallback\",\n   meta: {\n@@ -63,6 +72,21 @@ export const noErrStackThenStringFallbackRule = createRule({\n   },\n   defaultOptions: [],\n   create(context) {\n+    const sourceCode = context.sourceCode;\n+    type SourceCodeScope = ReturnType<typeof sourceCode.getScope>;\n+\n+    function hasResolvableLocalBinding(node: TSESTree.Node, name: string): boolean {\n+      let scope: SourceCodeScope | null = sourceCode.getScope(node);\n+      while (scope) {\n+        const variable = scope.set.get(name);\n+        if (variable && variable.defs.some(def => isDefinitionAvailableAtNode(def, node))) {\n+          return true;\n+        }\n+        scope = scope.upper;\n+      }\n+      return false;\n+    }\n+\n     return {\n       ConditionalExpression(node) {\n         // Patterns:\n@@ -91,15 +115,17 @@ export const noErrStackThenStringFallbackRule = createRule({\n           node,\n           messageId: \"preferGetErrorMessage\",\n           data: { errorVar: errVar },\n-          suggest: [\n-            {\n-              messageId: \"replaceWithGetErrorMessage\",\n-              data: { errorVar: errVar },\n-              fix(fixer) {\n-                return fixer.replaceText(node, `getErrorMessage(${errVar})`);\n-              },\n-            },\n-          ],\n+          suggest: hasResolvableLocalBinding(node, \"getErrorMessage\")\n+            ? [\n+                {\n+                  messageId: \"replaceWithGetErrorMessage\",\n+                  data: { errorVar: errVar },\n+                  fix(fixer) {\n+                    return fixer.replaceText(node, `getErrorMessage(${errVar})`);\n+                  },\n+                },\n+              ]\n+            : [],\n         });\n       },\n     };\n```\n\nAlso included in the diff (test file additions, and an unrelated generated lock.yml file change with pricing/model-routing config additions unrelated to this ESLint fix).\n\nReport findings as strict JSONL, one JSON object per line, each with fields: file, line (best estimate), severity (critical/high/medium/low), issue (short), rationale (why it matters). Do not include markdown, only JSONL output."}}}
[copilot-sdk-driver] [sdk-driver] error: Execution failed: Error: No model available. Check policy enablement under GitHub Settings > Copilot
{"type":"subagent.completed","timestamp":"2026-08-01T07:48:07.405Z","data":{"agentName":"general-purpose","toolCallId":"toolu_01DU3A9boP1zkM7asH8wxwqi"}}

Generated from PR Code Quality Reviewer · 58.6 AIC · ◷

[github-actions]

Agent job 30690731453 failed.

Warning

Engine Failure: The copilot engine terminated unexpectedly.

Last agent output:

{"type":"tool.execution_complete","timestamp":"2026-08-01T08:04:11.588Z","data":{"toolName":"bash","mcpServerName":"","success":true,"result":{"content":"import { describe, it, expect, beforeEach, vi } from \"vitest\";\nimport { readFileSync } from \"fs\";\nimport path from \"path\";\nconst createTestableFunction = scriptContent => {\n  const beforeMainCall = scriptContent.match(/^([\\s\\S]*?)\\s*module\\.exports\\s*=\\s*{[\\s\\S]*?};?\\s*$/);\n  if (!beforeMainCall) throw new Error(\"Could not extract script content before module.exports\");\n  let scriptBody = beforeMainCall[1];\n  // Mock the error_helpers and messages_core modules\n  const mockRequire = module => {\n    if (module === \"./error_helpers.cjs\") {\n      return { getErrorMessage: error => (error instanceof Error ? error.message : String(error)) };\n    }\n    if (module === \"./messages_core.cjs\") {\n      return {\n        getMessages: () => {\n          // Return messages config from environment\n          const messagesEnv = process.env.GH_AW_SAFE_OUTPUT_MESSAGES;\n          if (!messagesEnv) return null;\n          try {\n            return JSON.parse(messagesEnv);\n          } catch (e) {\n            return null;\n          }\n        },\n      };\n    }\n    if (module === \"./sanitize_content.cjs\") {\n      // Mock sanitizeContent to return input as-is for testing\n      return { sanitizeContent: content => content };\n    }\n    if (module === \"./messages_run_status.cjs\") {\n      const messagesEnv = process.env.GH_AW_SAFE_OUTPUT_MESSAGES;\n      const messages = messagesEnv ? JSON.parse(messagesEnv) : null;\n      const renderTmpl = (tmpl, ctx) => tmpl.replace(/\\{(\\w+)\\}/g, (m, k) => (ctx[k] !== undefined ? String(ctx[k]) : m));\n      const toSnake = obj => {\n        const r = {};\n        for (const [k, v] of Object.entries(obj)) {\n          r[k.replace(/([A-Z])/g, \"_$1\").toLowerCase()] = v;\n          r[k] = v;\n        }\n        return r;\n      };\n      return {\n        getPullRequestCreatedMessage: ctx => {\n          const tc = toSnake(ctx);\n          const def = \"Pull request created: [#{item_number}]({item_url})\";\n          return messages?.pullRequestCreated ? renderTmpl(messages.pullRequestCreated, tc) : renderTmpl(def, tc);\n        },\n        getIssueCreatedMessage: ctx => {\n          const tc = toSnake(ctx);\n          const def = \"Issue created: [#{item_number}]({item_url})\";\n          return messages?.issueCreated ? renderTmpl(messages.issueCreated, tc) : renderTmpl(def, tc);\n        },\n        getCommitPushedMessage: ctx => {\n          const tc = toSnake(ctx);\n          const def = \"Commit pushed: [`{short_sha}`]({commit_url})\";\n          return messages?.commitPushed ? renderTmpl(messages.commitPushed, tc) : renderTmpl(def, tc);\n        },\n      };\n    }\n    if (module === \"./templatable.cjs\") {\n      return {\n        parseBoolTemplatable: (value, defaultValue = true) => {\n          if (value === undefined || value === null) return defaultValue;\n          return String(value) !== \"false\";\n        },\n      };\n    }\n    if (module === \"./generate_footer.cjs\") {\n      return {\n        generateXMLMarker: (workflowName, runUrl) => `<!-- gh-aw-agentic-workflow: ${workflowName}, run: ${runUrl} -->`,\n      };\n    }\n    if (module === \"./messages_footer.cjs\") {\n      return {\n        getFooterMessage: ctx => `> Generated by [${ctx.workflowName}](${ctx.runUrl})`,\n        getDetectionCautionAlert: () => {\n          const conclusion = process.env.GH_AW_DETECTION_CONCLUSION;\n          if (conclusion !== \"warning\") return \"\";\n          const reason = process.env.GH_AW_DETECTION_REASON || \"\";\n          const reasons = {\n            threat_detected: \"Potential security threats were detected in the agent output.\",\n            agent_failure: \"The threat detection engine failed to produce results.\",\n            parse_error: \"The threat detection results could not be parsed.\",\n          };\n          const reasonText = reasons[reason] || \"The threat detection analysis could not be completed.\";\n          const isEngineError = reason === \"agent_failure\" || reason === \"parse_error\";\n          if (isEngineError) {\n            return `> [!WARNING]\\n> threat detection engine error\\n> The threat detection engine encountered an error and could not complete analysis. This is a tooling failure, not a security finding.\\n> <!-- gh-aw-threat-engine-error -->\\n>\\n> ${reasonText}`;\n          }\n          return `> [!CAUTION]\\n> agentic threat detected\\n> Threat detection flagged this output in warn mode. Manual review is REQUIRED before any follow-up automation.\\n> <!-- gh-aw-threat-detected -->\\n>\\n> ${reasonText}`;\n        },\n      };\n    }\n    if (module === \"./workflow_metadata_helpers.cjs\") {\n      return {\n        buildWorkflowRunUrl: (ctx, repo) => {\n          const server = ctx.serverUrl || \"https://github.com\";\n          return `${server}/${repo.owner}/${repo.repo}/actions/runs/${ctx.runId}`;\n        },\n      };\n    }\n    if (module === \"./github_api_helpers.cjs\") {\n      return {\n        resolveTopLevelDiscussionCommentId: async (githubClient, commentNodeId) => {\n          // In tests, simulate that the triggering comment is always a top-level comment\n          return commentNodeId;\n        },\n      };\n    }\n<shellId: 22 completed with exit code 0>","detailedContent":"import { describe, it, expect, beforeEach, vi } from \"vitest\";\nimport { readFileSync } from \"fs\";\nimport path from \"path\";\nconst createTestableFunction = scriptContent => {\n  const beforeMainCall = scriptContent.match(/^([\\s\\S]*?)\\s*module\\.exports\\s*=\\s*{[\\s\\S]*?};?\\s*$/);\n  if (!beforeMainCall) throw new Error(\"Could not extract script content before module.exports\");\n  let scriptBody = beforeMainCall[1];\n  // Mock the error_helpers and messages_core modules\n  const mockRequire = module => {\n    if (module === \"./error_helpers.cjs\") {\n      return { getErrorMessage: error => (error instanceof Error ? error.message : String(error)) };\n    }\n    if (module === \"./messages_core.cjs\") {\n      return {\n        getMessages: () => {\n          // Return messages config from environment\n          const messagesEnv = process.env.GH_AW_SAFE_OUTPUT_MESSAGES;\n          if (!messagesEnv) return null;\n          try {\n            return JSON.parse(messagesEnv);\n          } catch (e) {\n            return null;\n          }\n        },\n      };\n    }\n    if (module === \"./sanitize_content.cjs\") {\n      // Mock sanitizeContent to return input as-is for testing\n      return { sanitizeContent: content => content };\n    }\n    if (module === \"./messages_run_status.cjs\") {\n      const messagesEnv = process.env.GH_AW_SAFE_OUTPUT_MESSAGES;\n      const messages = messagesEnv ? JSON.parse(messagesEnv) : null;\n      const renderTmpl = (tmpl, ctx) => tmpl.replace(/\\{(\\w+)\\}/g, (m, k) => (ctx[k] !== undefined ? String(ctx[k]) : m));\n      const toSnake = obj => {\n        const r = {};\n        for (const [k, v] of Object.entries(obj)) {\n          r[k.replace(/([A-Z])/g, \"_$1\").toLowerCase()] = v;\n          r[k] = v;\n        }\n        return r;\n      };\n      return {\n        getPullRequestCreatedMessage: ctx => {\n          const tc = toSnake(ctx);\n          const def = \"Pull request created: [#{item_number}]({item_url})\";\n          return messages?.pullRequestCreated ? renderTmpl(messages.pullRequestCreated, tc) : renderTmpl(def, tc);\n        },\n        getIssueCreatedMessage: ctx => {\n          const tc = toSnake(ctx);\n          const def = \"Issue created: [#{item_number}]({item_url})\";\n          return messages?.issueCreated ? renderTmpl(messages.issueCreated, tc) : renderTmpl(def, tc);\n        },\n        getCommitPushedMessage: ctx => {\n          const tc = toSnake(ctx);\n          const def = \"Commit pushed: [`{short_sha}`]({commit_url})\";\n          return messages?.commitPushed ? renderTmpl(messages.commitPushed, tc) : renderTmpl(def, tc);\n        },\n      };\n    }\n    if (module === \"./templatable.cjs\") {\n      return {\n        parseBoolTemplatable: (value, defaultValue = true) => {\n          if (value === undefined || value === null) return defaultValue;\n          return String(value) !== \"false\";\n        },\n      };\n    }\n    if (module === \"./generate_footer.cjs\") {\n      return {\n        generateXMLMarker: (workflowName, runUrl) => `<!-- gh-aw-agentic-workflow: ${workflowName}, run: ${runUrl} -->`,\n      };\n    }\n    if (module === \"./messages_footer.cjs\") {\n      return {\n        getFooterMessage: ctx => `> Generated by [${ctx.workflowName}](${ctx.runUrl})`,\n        getDetectionCautionAlert: () => {\n          const conclusion = process.env.GH_AW_DETECTION_CONCLUSION;\n          if (conclusion !== \"warning\") return \"\";\n          const reason = process.env.GH_AW_DETECTION_REASON || \"\";\n          const reasons = {\n            threat_detected: \"Potential security threats were detected in the agent output.\",\n            agent_failure: \"The threat detection engine failed to produce results.\",\n            parse_error: \"The threat detection results could not be parsed.\",\n          };\n          const reasonText = reasons[reason] || \"The threat detection analysis could not be completed.\";\n          const isEngineError = reason === \"agent_failure\" || reason === \"parse_error\";\n          if (isEngineError) {\n            return `> [!WARNING]\\n> threat detection engine error\\n> The threat detection engine encountered an error and could not complete analysis. This is a tooling failure, not a security finding.\\n> <!-- gh-aw-threat-engine-error -->\\n>\\n> ${reasonText}`;\n          }\n          return `> [!CAUTION]\\n> agentic threat detected\\n> Threat detection flagged this output in warn mode. Manual review is REQUIRED before any follow-up automation.\\n> <!-- gh-aw-threat-detected -->\\n>\\n> ${reasonText}`;\n        },\n      };\n    }\n    if (module === \"./workflow_metadata_helpers.cjs\") {\n      return {\n        buildWorkflowRunUrl: (ctx, repo) => {\n          const server = ctx.serverUrl || \"https://github.com\";\n          return `${server}/${repo.owner}/${repo.repo}/actions/runs/${ctx.runId}`;\n        },\n      };\n    }\n    if (module === \"./github_api_helpers.cjs\") {\n      return {\n        resolveTopLevelDiscussionCommentId: async (githubClient, commentNodeId) => {\n          // In tests, simulate that the triggering comment is always a top-level comment\n          return commentNodeId;\n        },\n      };\n    }\n<shellId: 22 completed with exit code 0>"}}}
{"type":"assistant.message","timestamp":"2026-08-01T08:04:16.500Z","data":{"content":""}}
{"type":"tool.execution_start","timestamp":"2026-08-01T08:04:16.502Z","data":{"toolName":"bash","mcpServerName":"","command":"cd /home/runner/work/gh-aw/gh-aw && grep -n \"getDetectionCautionAlert\" -A 20 actions/setup/js/messages_footer.cjs | head -40"}}
{"type":"tool.execution_complete","timestamp":"2026-08-01T08:04:16.571Z","data":{"toolName":"bash","mcpServerName":"","success":true,"result":{"content":"28:function getDetectionCautionAlert(workflowName, runUrl) {\n29-  const detectionConclusion = process.env.GH_AW_DETECTION_CONCLUSION;\n30-  if (detectionConclusion !== \"warning\") {\n31-    return \"\";\n32-  }\n33-  const detectionReason = process.env.GH_AW_DETECTION_REASON || \"\";\n34-  return getDetectionWarningMessage({ workflowName, runUrl, reason: detectionReason });\n35-}\n36-\n37-/**\n38- * `@param` {string|undefined} raw\n39- * `@returns` {number|undefined}\n40- */\n41-function parsePositiveAIC(raw) {\n42-  const parsed = raw ? Number.parseFloat(raw) : NaN;\n43-  return Number.isFinite(parsed) && parsed > 0 ? parsed : undefined;\n44-}\n45-\n46-/**\n47- * `@param` {number|string|undefined} raw\n48- * `@returns` {number|undefined}\n--\n714:    const detectionCaution = getDetectionCautionAlert(workflowName, runUrl);\n715-    if (detectionCaution) {\n716-      guardNotices += detectionCaution;\n717-    }\n718-  }\n719-\n720-  // Add firewall blocked domains section if any domains were blocked\n721-  const blockedDomains = getBlockedDomains();\n722-  const blockedDomainsSection = generateBlockedDomainsSection(blockedDomains);\n723-  if (blockedDomainsSection) {\n724-    guardNotices += blockedDomainsSection;\n725-  }\n726-\n727-  // Add integrity filtering section if any items were filtered\n728-  try {\n729-    const difcFilteredEvents = getDifcFilteredEvents();\n730-    const difcFilteredSection = generateDifcFilteredSection(difcFilteredEvents);\n731-    if (difcFilteredSection) {\n<shellId: 23 completed with exit code 0>","detailedContent":"28:function getDetectionCautionAlert(workflowName, runUrl) {\n29-  const detectionConclusion = process.env.GH_AW_DETECTION_CONCLUSION;\n30-  if (detectionConclusion !== \"warning\") {\n31-    return \"\";\n32-  }\n33-  const detectionReason = process.env.GH_AW_DETECTION_REASON || \"\";\n34-  return getDetectionWarningMessage({ workflowName, runUrl, reason: detectionReason });\n35-}\n36-\n37-/**\n38- * `@param` {string|undefined} raw\n39- * `@returns` {number|undefined}\n40- */\n41-function parsePositiveAIC(raw) {\n42-  const parsed = raw ? Number.parseFloat(raw) : NaN;\n43-  return Number.isFinite(parsed) && parsed > 0 ? parsed : undefined;\n44-}\n45-\n46-/**\n47- * `@param` {number|string|undefined} raw\n48- * `@returns` {number|undefined}\n--\n714:    const detectionCaution = getDetectionCautionAlert(workflowName, runUrl);\n715-    if (detectionCaution) {\n716-      guardNotices += detectionCaution;\n717-    }\n718-  }\n719-\n720-  // Add firewall blocked domains section if any domains were blocked\n721-  const blockedDomains = getBlockedDomains();\n722-  const blockedDomainsSection = generateBlockedDomainsSection(blockedDomains);\n723-  if (blockedDomainsSection) {\n724-    guardNotices += blockedDomainsSection;\n725-  }\n726-\n727-  // Add integrity filtering section if any items were filtered\n728-  try {\n729-    const difcFilteredEvents = getDifcFilteredEvents();\n730-    const difcFilteredSection = generateDifcFilteredSection(difcFilteredEvents);\n731-    if (difcFilteredSection) {\n<shellId: 23 completed with exit code 0>"}}}
{"type":"assistant.message","timestamp":"2026-08-01T08:04:31.109Z","data":{"content":""}}
{"type":"tool.execution_start","timestamp":"2026-08-01T08:04:31.111Z","data":{"toolName":"task","mcpServerName":""}}
{"type":"subagent.started","timestamp":"2026-08-01T08:04:31.121Z","data":{"agentName":"general-purpose","agentDisplayName":"General Purpose Agent","toolCallId":"toolu_01CCZWr36zozCGX71STyR2Qn"}}
[copilot-sdk-driver] [sdk-driver] error: Execution failed: Error: No model available. Check policy enablement under GitHub Settings > Copilot
{"type":"subagent.completed","timestamp":"2026-08-01T08:04:31.164Z","data":{"agentName":"general-purpose","toolCallId":"toolu_01CCZWr36zozCGX71STyR2Qn"}}
{"type":"tool.execution_complete","timestamp":"2026-08-01T08:04:31.189Z","data":{"toolName":"task","mcpServerName":"","success":false}}

Generated from PR Code Quality Reviewer · 83.2 AIC · ◷

[github-actions]

This issue was automatically closed because it expired on 2026-08-01T18:30:14.179Z.

Closed by Workflow