Dashboard › cli › Distillation
5d3d4b76-6608-4626-8c00-381dc1eb6c68["lore_tm_v1_IABT0pg8-ZSNlXWupprtkLdEUvRJD6U4OGtfn-EiggE","lore_tm_v1_Uj7IArz4uMRr-a5dJwDjEN-3ptquTlaSmEkjNy9sQqg","lore_tm_v1_iDEwZLgeHdjNe0_sevZGQz-QWeUq6UwrH4R8V9d5A3c","lore_tm_v1_hC1gSKAgRoo_Mckm-Q6X03J8sxDqxOKythEXx58dEVk","lore_tm_v1_omO2TCXZlWcpK4OLEld7NMwLuLrri7aEsKYJpLOBY1w","lore_tm_v1_VMHk571CAE2Xgd6V9_BCVB_0FYJ4kRbklnmVKfoTkgU","lore_tm_v1_RZMqvcZKa-7KjRPU5d9eNj5M0MvpuTVVB8YLySZroXg","lore_tm_v1_0qrHUpsFhWaajPmZVEIp0Dsqaa4Gtxig1A4PQW--Wlk","lore_tm_v1_koYbiSXkRaDHbuq3C36vSD10EAiKm8Koy9Q3iYdSMVQ","lore_tm_v1_es2oDmbKQW8Ud0_silcsfRMtR6pbflP1poEUaeNBqdo"]
packages/cli/script/check-fragments.ts now skips fragment subcommand-coverage checks for every command whose cmd.examples.length > 0; structured command metadata examples count as documentation, so only commands without metadata examples must be mentioned in the hand-written route fragment.packages/cli/script/generate-command-docs.ts now imports formatCommandExamples from ./generate-skill-markdown.js and includes each command’s canonical metadata examples in its generated reference section after the options table.docs/src/fragments/; fragments remain the source of truth only for content that cannot live in command definitions.formatPositionalsTable(positionals) in packages/cli/script/generate-command-docs.ts now renders optional positionals as [<placeholder>] and required positionals as <placeholder>.packages/cli/script/generate-skill-markdown.ts adds exported formatCommandExamples(examples: readonly string[]): string; an empty array returns "", while nonempty input renders **Examples:**, a fenced bash block, and examples separated by one blank line via examples.join("\n\n").packages/cli/script/generate-skill.ts now reuses formatCommandExamples(cmd.examples) in generateFullCommandDoc() instead of maintaining a separate examples-rendering implementation.packages/cli/test/script/generate-skill-markdown.test.ts adds an integration-style assertion that agent-conversation documentation comes from command metadata by introspecting listCommand and viewCommand with buildCommandInfo().listCommand.__primaryUsage to be undefined while its introspected positional signature is [<org>].sentry agent-conversation list examples are ordered as: 1. # List recent agent conversations\nsentry agent-conversation list; 2. # Explicit organization\nsentry agent-conversation list my-org; 3. # Show more, last 24 hours\nsentry agent-conversation list --limit 50 --period 24h; 4. # Filter conversations\nsentry agent-conversation list -q "has:errors"; 5. # Paginate through results\nsentry agent-conversation list my-org -c next.sentry agent-conversation view positional signature is [<org>/]<conversation-id>.sentry agent-conversation view examples are ordered exactly as: 1. # View full transcript (organization auto-detected)\nsentry agent-conversation view conv-123; 2. # Explicit organization\nsentry agent-conversation view my-org/conv-123; 3. # JSON output\nsentry agent-conversation view my-org/conv-123 --json.formatCommandExamples(view.examples) produces an **Examples:** section with a fenced bash block, preserves each # <description> immediately above its command, and inserts a blank line between examples.sentry agent-conversation view native help has brief View an agent conversation transcript, primary usage [<org>/]<conversation-id>, and explains that the organization is optional, auto-detected from project context when omitted, and may be supplied by prefixing the conversation ID with an organization slug.sentry agent-conversation list native help has brief List recent agent conversations, states that the organization is auto-detected when omitted, and uses one optional tuple parameter with placeholder org, brief Organization slug, and optional: true.sentry agent-conversation list documents 17 JSON fields: conversationId (string), title (string | null, optional), flow (array), errors (number), llmCalls (number), toolCalls (number), totalTokens (number), totalCost (number), startTimestamp (number), endTimestamp (number), traceCount (number), traceIds (array), firstInput (string | null), lastOutput (string | null), user (object | null, optional), toolNames (array), and toolErrors (number).sentry project create arguments to always be attached to the project name with :; space-separated forms are invalid.sentry project create "My New App":, sentry project create my-new-app javascript-nextjs, and sentry project create my-org/my-new-app javascript-nextjs; it requires examples including sentry project create web:javascript api:python-django worker:node and sentry project create my-new-app:javascript-nextjs.extractCommandPathFromHeading() is tested against the optional/variadic heading sentry project create [<org>/]<name>:<platform>... and must extract the literal path sentry project create.SentryContext.commandPrefix in packages/cli/src/context.ts is an optional readonly array containing the full command prefix including the program name, such as ["sentry", "issue", "list"]; buildContext(...).forCommand() stores Stricli’s prefix there and sets the telemetry span name with prefix.join(".").packages/cli/src/cli.ts is best-effort: when Stricli sets ExitCode.UnknownCommand (including its modulo-256 representation), recoverable help/version requests are excluded, reportUnknownCommand(argv) is attempted, and telemetry failures are silently caught so they cannot crash command execution.packages/cli/src/app.ts configures the application name as sentry, scanner caseStyle: "allow-kebab-for-camel", allowArgumentEscapeSequence: true, top-level global flags from buildTopLevelFlags(), exit-code determination through getExitCode, and JSON help through renderJsonHelp(prefix, unprocessedInputs).