@sanity/ailf 3.7.0 → 3.8.0

package/dist/commands/{pipeline.js → run.js}

@@ -1,23 +1,29 @@
 /**
- * pipeline command — the main evaluation pipeline orchestrator.
+ * run command — the main evaluation entry point.
  *
- * Defines all 36+ CLI flags via Commander, resolves them into a typed
+ * Renamed from `ailf pipeline` to `ailf run` per W0077 Phase 1
+ * (see docs/design-docs/pipeline-command-surface.md §3). The command drives
+ * the evaluation pipeline; the orchestrator internals retain the "pipeline"
+ * name because they describe a multi-step process, not the CLI verb.
+ *
+ * Defines all 44+ CLI flags via Commander, resolves them into a typed
  * options object, bridges to process.env for downstream modules, and
  * delegates to runPipeline().
  *
- * @see docs/cli.md for the full flag reference.
+ * @see docs/guides/cli-guide.md for per-flag behavior notes, or
+ * docs/references/cli-reference.md for the auto-generated flag matrix.
  */
 import { Command } from "commander";
-import { addAgenticOptions, addDebugOptions, addSanitySourceOptions, } from "./shared/options.js";
-export function createPipelineCommand() {
-    const cmd = new Command("pipeline")
+import { addAgenticOptions, addDebugOptions, addSanityScopeOptions, } from "./shared/options.js";
+export function createRunCommand() {
+    const cmd = new Command("run")
         .description("Run the full evaluation pipeline")
         .option("-m, --mode <mode>", "Evaluation mode: literacy (default), mcp-server, agent-harness, knowledge-probe, custom. Legacy aliases (baseline, agentic, observed, full) are accepted and normalized to literacy + variant.", "literacy")
         .option("--variant <variant>", "Literacy variant: full (default — standard + agentic), baseline (standard only), agentic (agentic only), observed. Only applies to --mode literacy.")
         .option("-s, --source <name>", "Documentation source name (from sources.yaml)")
         .option("-n, --dry-run", "Validate configuration only, no execution", false)
-        .option("--skip-fetch", "Reuse cached documentation contexts", false)
-        .option("--skip-eval", "Recalculate from existing eval results", false)
+        .option("--no-fetch", "Reuse cached documentation contexts")
+        .option("--no-eval", "Recalculate from existing eval results")
         .option("--no-cache", "Bypass all pipeline-level caching")
         .option("--no-remote-cache", "Disable Content Lake cache lookup (local cache still active)")
         .option("--no-auto-scope", "Disable release-aware auto-scoping (evaluate all tasks even when a perspective is set)")
@@ -31,39 +37,22 @@ export function createPipelineCommand() {
             .filter(Boolean),
     ], [])
         .option("--changed-docs <slugs>", "Auto-scope to tasks affected by these document slugs")
-        .option("-j, --concurrency <n>", "Max parallel API calls during evaluation", parseInt)
-        .option("--grader-replications <n>", "Grader consistency replications", parseInt)
-        .option("--before <source>", "Before-state for impact evaluation")
-        .option("-c, --compare", "Compare scores against latest baseline", false)
-        .option("--compare-baseline <path>", "Specific baseline file to compare")
+        .option("--before-source <name>", "Swap the doc source to a `before` state and run a paired evaluation for impact analysis. Pairs with `ailf baseline` and `--compare`. Distinct from `--compare <baseline>`, which compares scores against a saved snapshot.")
+        .option("-c, --compare [baseline]", "Compare scores against the latest baseline. Pass a path to pin a specific baseline file (e.g. --compare results/baselines/2026-04-22.json).")
         .option("--threshold <n>", "Noise threshold for comparison (default: 2)", parseFloat)
-        .option("--no-gap-analysis", "Skip failure mode + impact analysis")
-        .option("--readiness", "Generate launch readiness checklist", false)
-        .option("--discovery-report", "Generate agent discoverability report", false)
         .option("-p, --publish", "Write report to Sanity + fan out to sinks (auto-enabled for full runs when report store is configured)")
         .option("--no-publish", "Suppress auto-publishing")
         .option("--publish-tag <tag>", "Label for published report")
-        .option("--report-dataset <name>", "Sanity dataset for report store")
-        .option("--report-project <id>", "Sanity project ID for report store")
         .option("--config <path>", "Load pipeline config from a TS/JS/YAML/JSON file (overrides most CLI flags)")
         .option("-o, --output <path>", "Write PR comment markdown to file")
-        .option("--output-dir <path>", "Base directory for pipeline output artifacts (default: inferred from execution context)")
         .option("--promptfoo-url <url>", "Promptfoo share URL for report")
-        .option("--task-source <type>", "Task definition source: content-lake (default — Sanity Content Lake), repo (repo tasks only, no Content Lake merge)", "content-lake")
-        .option("--repo-tasks-path <path>", "Path to repo-based task definitions. Defaults to ./.ailf/tasks/ when --task-source=repo.")
         .option("--remote", "Submit evaluation to the AILF API instead of running locally", false)
-        .option("--api-url <url>", "AILF API base URL (default: https://ailf-api.sanity.build)")
-        .option("--no-artifacts", "Disable all artifact writers (D0033). Overrides --artifacts-dir.")
-        .option("--artifacts-dir <path>", "Root directory for local artifact output (D0033; default: .ailf/results/captures/)")
-        .option("--artifacts-dry-run", "Run artifact writers in dry-run mode — log intended writes, touch no storage", false)
-        .option("--artifacts-exclude <types>", "Comma-separated artifact types to skip (e.g. traces,graderPrompts)")
+        .option("--no-artifacts-write", "Run artifact writers in dry-run mode — log intended writes, touch no storage")
         // D0037 caller envelope (W0069) — threads through --remote so the
         // server-side pipeline attributes provenance to the caller, not the
         // API gateway runner. All env-var equivalents are honored too;
         // explicit flags win over env vars.
-        .option("--classification <value>", "Run classification for provenance: official | ad-hoc | experimental | test | external. Overrides AILF_CLASSIFICATION. See D0037.")
-        .option("--owner-team <slug>", "Team slug this run is attributable to. Overrides AILF_OWNER_TEAM.")
-        .option("--owner-individual <slug>", "Individual (GH actor / user ID) this run is attributable to. Overrides AILF_OWNER_INDIVIDUAL.")
+        .option("--classification <value>", "Run classification for provenance: official | adhoc | experimental | test | external. Overrides AILF_CLASSIFICATION. See D0037.")
         .option("--purpose <text>", 'Free-text "why I ran this" attached to provenance. Overrides AILF_PURPOSE.')
         .option("--label <value>", "Free-form searchable label (repeatable). Appends to any AILF_LABELS env value.", (val, prev) => [
         ...prev,
@@ -78,7 +67,7 @@ export function createPipelineCommand() {
     });
     // Add shared option groups
     addDebugOptions(cmd);
-    addSanitySourceOptions(cmd);
+    addSanityScopeOptions(cmd);
     addAgenticOptions(cmd);
     return cmd;
 }

package/dist/commands/shared/help.js

@@ -67,14 +67,14 @@ function hasColorSupport() {
 // ---------------------------------------------------------------------------
 const afterHelpText = `
 Quick Start:
-  $ ailf pipeline --debug          Run a quick evaluation (first 2 tests)
-  $ ailf pipeline --area groq      Evaluate a specific feature area
-  $ ailf pipeline --explain        Preview the execution plan
+  $ ailf run --debug               Run a quick evaluation (first 2 tests)
+  $ ailf run --area groq           Evaluate a specific feature area
+  $ ailf run --explain             Preview the execution plan
   $ ailf init                      Set up AILF in a new project
 
 Documentation:
   Repository   https://github.com/sanity-io/ai-literacy-framework
-  CLI Guide    https://github.com/sanity-io/ai-literacy-framework/blob/main/docs/cli.md
+  CLI Guide    https://github.com/sanity-io/ai-literacy-framework/blob/main/docs/guides/cli-guide.md
   Getting Started  https://github.com/sanity-io/ai-literacy-framework/blob/main/docs/getting-started.md
 
 Run ailf <command> --help for detailed usage of any command.`;

package/dist/commands/shared/options.d.ts

@@ -7,11 +7,22 @@
  */
 import type { Command } from "commander";
 /**
- * Add agentic options: --url, --header, --allowed-origin, --search
+ * Add agentic options on `ailf run`: `--url` and `--search`. Both are
+ * per-run overrides.
+ *
+ * **`--header` and `--allowed-origin` retired in W0077 Phase 6f** —
+ * configure them in `.ailf/config.yaml`'s `agentic` block instead. The
+ * `DOC_HEADERS` and `DOC_ALLOWED_ORIGIN(S)` env vars still apply.
  */
 export declare function addAgenticOptions(cmd: Command): Command;
 /**
- * Add debug options: --debug, --debug-n, --debug-pattern, --debug-sample
+ * Add debug + filter options.
+ *
+ * `-d, --debug` is a shortcut that runs only the first 2 tests for fast
+ * feedback. `--filter-first-n`, `--filter-pattern`, and `--filter-sample`
+ * narrow the test set explicitly and mirror the same-named flags on
+ * `ailf eval` so the pipeline surface matches the underlying Promptfoo
+ * passthrough.
  */
 export declare function addDebugOptions(cmd: Command): Command;
 /**
@@ -27,9 +38,24 @@ export declare function addOutputOptions(cmd: Command): Command;
  */
 export declare function addOutputDirOption(cmd: Command): Command;
 /**
- * Add Sanity source options: --sanity-dataset, --sanity-project, etc.
+ * Add the full Sanity-source CLI surface — dataset, project, perspective,
+ * studio origin, document. Used by `ailf fetch-docs`, which is invoked
+ * directly with explicit Sanity targeting per run.
+ *
+ * **Not used by `ailf run`** post-W0077 Phase 6d. The dataset, project, and
+ * studio-origin trio moved to `.ailf/config.yaml`'s `source` block (with
+ * `SANITY_DATASET` / `SANITY_PROJECT_ID` / `SANITY_STUDIO_ORIGIN` env-var
+ * fallbacks). `ailf run` uses `addSanityScopeOptions` instead, which keeps
+ * only the per-run flags (`--sanity-perspective`, `--sanity-document`).
  */
 export declare function addSanitySourceOptions(cmd: Command): Command;
+/**
+ * Add the per-run Sanity-scope CLI surface — `--sanity-perspective` and
+ * `--sanity-document`. Used by `ailf run`. The per-environment trio
+ * (dataset, project, studio origin) lives in `.ailf/config.yaml`'s
+ * `source` block instead (W0077 Phase 6d).
+ */
+export declare function addSanityScopeOptions(cmd: Command): Command;
 /**
  * Collect repeatable string options into an array.
  * Used as a Commander argParser for options like --url, --header, --allowed-origin.

package/dist/commands/shared/options.js

@@ -6,27 +6,33 @@
  * (e.g., debug options, Sanity source options, output options).
  */
 /**
- * Add agentic options: --url, --header, --allowed-origin, --search
+ * Add agentic options on `ailf run`: `--url` and `--search`. Both are
+ * per-run overrides.
+ *
+ * **`--header` and `--allowed-origin` retired in W0077 Phase 6f** —
+ * configure them in `.ailf/config.yaml`'s `agentic` block instead. The
+ * `DOC_HEADERS` and `DOC_ALLOWED_ORIGIN(S)` env vars still apply.
  */
 export function addAgenticOptions(cmd) {
     return cmd
         .option("--url <url>", "Documentation URL (repeatable)", collect, [])
-        .option("--urls <url>", "Alias for --url (repeatable)", collect, [])
-        .option("--header <header>", 'Custom HTTP header "Key: Value" (repeatable)', collect, [])
-        .option("--headers <header>", "Alias for --header (repeatable)", collect, [])
-        .option("--allowed-origin <origin>", "Agent origin sandbox (repeatable, supports globs)", collect, [])
-        .option("--allowed-origins <origin>", "Alias for --allowed-origin (repeatable)", collect, [])
         .option("-S, --search <mode>", "Web search mode: open, origin-only, off");
 }
 /**
- * Add debug options: --debug, --debug-n, --debug-pattern, --debug-sample
+ * Add debug + filter options.
+ *
+ * `-d, --debug` is a shortcut that runs only the first 2 tests for fast
+ * feedback. `--filter-first-n`, `--filter-pattern`, and `--filter-sample`
+ * narrow the test set explicitly and mirror the same-named flags on
+ * `ailf eval` so the pipeline surface matches the underlying Promptfoo
+ * passthrough.
  */
 export function addDebugOptions(cmd) {
     return cmd
         .option("-d, --debug", "Run subset of tests for fast feedback", false)
-        .option("--debug-n <n>", "First N tests", parseInt)
-        .option("--debug-pattern <regex>", "Filter tests by description regex")
-        .option("--debug-sample <n>", "Random sample of N tests", parseInt);
+        .option("--filter-first-n <n>", "Run only first N tests", parseInt)
+        .option("--filter-pattern <regex>", "Filter tests by description regex")
+        .option("--filter-sample <n>", "Random sample of N tests", parseInt);
 }
 /**
  * Add output options: --output, --format
@@ -47,7 +53,15 @@ export function addOutputDirOption(cmd) {
     return cmd.option("--output-dir <path>", "Base directory for output artifacts (default: .ailf/results/latest/)");
 }
 /**
- * Add Sanity source options: --sanity-dataset, --sanity-project, etc.
+ * Add the full Sanity-source CLI surface — dataset, project, perspective,
+ * studio origin, document. Used by `ailf fetch-docs`, which is invoked
+ * directly with explicit Sanity targeting per run.
+ *
+ * **Not used by `ailf run`** post-W0077 Phase 6d. The dataset, project, and
+ * studio-origin trio moved to `.ailf/config.yaml`'s `source` block (with
+ * `SANITY_DATASET` / `SANITY_PROJECT_ID` / `SANITY_STUDIO_ORIGIN` env-var
+ * fallbacks). `ailf run` uses `addSanityScopeOptions` instead, which keeps
+ * only the per-run flags (`--sanity-perspective`, `--sanity-document`).
  */
 export function addSanitySourceOptions(cmd) {
     return cmd
@@ -55,8 +69,18 @@ export function addSanitySourceOptions(cmd) {
         .option("--sanity-project <id>", "Override Sanity project ID")
         .option("--sanity-perspective <id>", "Sanity release perspective ID")
         .option("--sanity-studio-origin <url>", "Sanity Studio base URL")
-        .option("--sanity-document <id>", "Evaluate specific Sanity document(s) (repeatable)", collect, [])
-        .option("--sanity-documents <id>", "Alias for --sanity-document (repeatable)", collect, []);
+        .option("--sanity-document <id>", "Evaluate specific Sanity document(s) (repeatable)", collect, []);
+}
+/**
+ * Add the per-run Sanity-scope CLI surface — `--sanity-perspective` and
+ * `--sanity-document`. Used by `ailf run`. The per-environment trio
+ * (dataset, project, studio origin) lives in `.ailf/config.yaml`'s
+ * `source` block instead (W0077 Phase 6d).
+ */
+export function addSanityScopeOptions(cmd) {
+    return cmd
+        .option("--sanity-perspective <id>", "Sanity release perspective ID")
+        .option("--sanity-document <id>", "Evaluate specific Sanity document(s) (repeatable)", collect, []);
 }
 /**
  * Collect repeatable string options into an array.

package/dist/commands/validate-tasks.js

@@ -21,7 +21,7 @@ import { detectLegacyFieldNames, parseCanonicalTaskFile, } from "../adapters/tas
 import { validateCanonicalTasks, formatValidationResult, } from "../adapters/task-sources/repo-validation.js";
 import { discoverTsTaskFiles, loadTsTaskFile, } from "../adapters/task-sources/task-file-loader.js";
 export function createValidateTasksCommand() {
-    return new Command("validate-tasks")
+    return new Command("tasks")
         .description("Validate task files (YAML and TypeScript) in .ailf/tasks/ against the canonical schema")
         .argument("[path]", "Path to tasks directory (default: .ailf/tasks/)", ".ailf/tasks")
         .option("--strict", "Treat warnings as errors", false)

package/dist/commands/validate.d.ts

@@ -6,4 +6,4 @@
  *
  */
 import { Command } from "commander";
-export declare function createValidateCommand(): Command;
+export declare function createValidateConfigCommand(): Command;

package/dist/commands/validate.js

@@ -10,8 +10,8 @@ import { dirname, resolve } from "path";
 import { fileURLToPath } from "url";
 const __dirname = dirname(fileURLToPath(import.meta.url));
 const ROOT = resolve(__dirname, "..", "..");
-export function createValidateCommand() {
-    return new Command("validate")
+export function createValidateConfigCommand() {
+    return new Command("config")
         .description("Validate pipeline configuration")
         .option("--strict", "Treat warnings as errors", false)
         .option("--contexts", "Check that context files exist", false)

package/dist/commands/weekly-digest.js

@@ -8,11 +8,11 @@
  */
 import { Command } from "commander";
 export function createWeeklyDigestCommand() {
-    return new Command("weekly-digest")
+    return new Command("digest")
         .description("Generate and deliver a weekly evaluation digest")
         .option("-n, --dry-run", "Print to stdout only, do not send to Slack", false)
         .option("--lookback <days>", "Lookback window in days (default: from config or 7)", parseInt)
-        .option("--json", "Output raw JSON digest data", false)
+        .option("-f, --format <fmt>", "Output format: console or json", "console")
         .action(async (opts) => {
         const { generateDigest } = await import("../schedules/digest.js");
         const { getDigestConfig } = await import("../schedules/loader.js");
@@ -45,7 +45,7 @@ export function createWeeklyDigestCommand() {
         console.log(`  Regressed: ${digest.regressed.join(", ") || "none"}`);
         console.log(`  Stable: ${digest.stable.join(", ") || "none"}`);
         console.log();
-        if (opts.json) {
+        if (opts.format === "json") {
             console.log(JSON.stringify(digest, null, 2));
             return;
         }

package/dist/config/thresholds.ts

@@ -2,9 +2,9 @@
  * thresholds.ts — Quality thresholds for readiness gates and regression alerts.
  *
  * Used by:
- * - `npx @sanity/ailf pipeline --readiness` (launch readiness checklist)
- * - `npx @sanity/ailf pipeline --publish` (severity-aware sink routing)
- * - `npx @sanity/ailf pipeline --compare` (regression alerting)
+ * - `npx @sanity/ailf report readiness` (launch readiness checklist)
+ * - `npx @sanity/ailf run --publish` (severity-aware sink routing)
+ * - `npx @sanity/ailf run --compare` (regression alerting)
  *
  * @see docs/archive/exec-plans/scenario-matrix-implementation/phase-5-readiness-thresholds.md
  */

package/dist/orchestration/build-app-context.js

@@ -44,8 +44,6 @@ export function mapToResolvedConfig(opts, rootDir) {
         compareThreshold: opts.compareThreshold,
         compareBaseline: opts.compareBaseline,
         gapAnalysisEnabled: opts.gapAnalysisEnabled,
-        readinessEnabled: opts.readinessEnabled,
-        discoveryReportEnabled: opts.discoveryReportEnabled,
         publishEnabled: opts.publishEnabled,
         publishTag: opts.publishTag,
         noCache: opts.noCache,

package/dist/orchestration/build-step-sequence.js

@@ -9,14 +9,12 @@ import { LiteracyVariant } from "../pipeline/normalize-mode.js";
 import { CallbackStep } from "./steps/callback-step.js";
 import { CalculateScoresStep } from "./steps/calculate-scores-step.js";
 import { CompareStep } from "./steps/compare-step.js";
-import { DiscoveryReportStep } from "./steps/discovery-report-step.js";
 import { FetchDocsStep } from "./steps/fetch-docs-step.js";
 import { FinalizeRunStep } from "./steps/finalize-run-step.js";
 import { GapAnalysisStep } from "./steps/gap-analysis-step.js";
 import { GenerateConfigsStep } from "./steps/generate-configs-step.js";
 import { GraderConsistencyStep } from "./steps/grader-consistency-step.js";
 import { PublishReportStep } from "./steps/publish-report-step.js";
-import { ReadinessStep } from "./steps/readiness-step.js";
 import { ReportStep } from "./steps/report-step.js";
 import { RunEvalStep } from "./steps/run-eval-step.js";
 import { MirrorRepoTasksStep } from "./steps/mirror-repo-tasks-step.js";
@@ -93,15 +91,7 @@ export function buildStepSequence(ctx, pipelineStart = Date.now()) {
     if (config.compareEnabled) {
         steps.push(new CompareStep());
     }
-    // Step 6b: Readiness report (optional)
-    if (config.readinessEnabled) {
-        steps.push(new ReadinessStep());
-    }
-    // Step 6c: Discovery report (optional)
-    if (config.discoveryReportEnabled) {
-        steps.push(new DiscoveryReportStep());
-    }
-    // Step 7: Callback delivery (optional, API-triggered evaluations)
+    // Step 6: Callback delivery (optional, API-triggered evaluations)
     if (config.callback?.url) {
         steps.push(new CallbackStep(config.callback, config.jobId));
     }

package/dist/orchestration/steps/fetch-docs-step.js

@@ -27,7 +27,7 @@ export class FetchDocsStep {
     }
     async execute(ctx, state) {
         if (ctx.config.skipFetch) {
-            return { status: "skipped", reason: "--skip-fetch" };
+            return { status: "skipped", reason: "--no-fetch" };
         }
         const start = Date.now();
         // Load tasks — use the same source as GenerateConfigsStep to avoid

package/dist/orchestration/steps/index.d.ts

@@ -6,14 +6,12 @@
  */
 export { CalculateScoresStep } from "./calculate-scores-step.js";
 export { CompareStep } from "./compare-step.js";
-export { DiscoveryReportStep } from "./discovery-report-step.js";
 export { FetchDocsStep } from "./fetch-docs-step.js";
 export { GapAnalysisStep } from "./gap-analysis-step.js";
 export { GenerateConfigsStep } from "./generate-configs-step.js";
 export { MirrorRepoTasksStep } from "./mirror-repo-tasks-step.js";
 export { GraderConsistencyStep } from "./grader-consistency-step.js";
 export { PublishReportStep } from "./publish-report-step.js";
-export { ReadinessStep } from "./readiness-step.js";
 export { ReportStep } from "./report-step.js";
 export { RunEvalStep } from "./run-eval-step.js";
 export { ValidateStep } from "./validate-step.js";

package/dist/orchestration/steps/index.js

@@ -6,14 +6,12 @@
  */
 export { CalculateScoresStep } from "./calculate-scores-step.js";
 export { CompareStep } from "./compare-step.js";
-export { DiscoveryReportStep } from "./discovery-report-step.js";
 export { FetchDocsStep } from "./fetch-docs-step.js";
 export { GapAnalysisStep } from "./gap-analysis-step.js";
 export { GenerateConfigsStep } from "./generate-configs-step.js";
 export { MirrorRepoTasksStep } from "./mirror-repo-tasks-step.js";
 export { GraderConsistencyStep } from "./grader-consistency-step.js";
 export { PublishReportStep } from "./publish-report-step.js";
-export { ReadinessStep } from "./readiness-step.js";
 export { ReportStep } from "./report-step.js";
 export { RunEvalStep } from "./run-eval-step.js";
 export { ValidateStep } from "./validate-step.js";

package/dist/orchestration/steps/run-eval-step.js

@@ -25,7 +25,7 @@ export class RunEvalStep {
     }
     async execute(ctx, state) {
         if (ctx.config.skipEval) {
-            return { status: "skipped", reason: "--skip-eval" };
+            return { status: "skipped", reason: "--no-eval" };
         }
         const start = Date.now();
         const { rootDir, debug, concurrency, noCache } = ctx.config;

package/dist/pipeline/cache.d.ts

@@ -48,7 +48,7 @@ export interface CacheStats {
     hits: number;
     /** Steps where cache was missed (executed normally) */
     misses: number;
-    /** Steps that were skipped for other reasons (--skip-fetch, etc.) */
+    /** Steps that were skipped for other reasons (--no-fetch, etc.) */
     skipped: number;
     /** Per-step detail */
     steps: Record<string, "disabled" | "hit" | "miss" | "skipped">;

package/dist/pipeline/map-request-to-config.js

@@ -47,8 +47,6 @@ export function mapRequestToConfig(request, rootDir) {
         compareThreshold: request.compareThreshold,
         compareBaseline: request.compareBaseline,
         gapAnalysisEnabled: request.gapAnalysis ?? true,
-        readinessEnabled: request.readiness ?? false,
-        discoveryReportEnabled: request.discoveryReport ?? false,
         publishEnabled: request.publish ?? publishDefault,
         publishTag: request.publishTag,
         noAutoScope: request.noAutoScope ?? false,

package/dist/pipeline/plan.d.ts

@@ -61,7 +61,7 @@ export interface DebugPlan {
 export interface ExecutionPlan {
     /** Cache prediction per step */
     cacheStatus: Record<string, "hit" | "miss" | "skipped" | "unknown">;
-    /** Command name (e.g., "pipeline", "compare", "validate") */
+    /** Command name (e.g., "run", "compare", "validate") */
     command: string;
     /** Comparison plan (when --compare is set) */
     comparison?: ComparisonPlan;
@@ -134,7 +134,6 @@ export interface PlanOptions {
     compareThreshold?: number;
     concurrency?: number;
     debug?: DebugOptions;
-    discoveryReportEnabled: boolean;
     dryRun: boolean;
     gapAnalysisEnabled: boolean;
     graderReplications?: number;
@@ -143,7 +142,6 @@ export interface PlanOptions {
     variant?: string;
     noCache: boolean;
     publishEnabled: boolean;
-    readinessEnabled: boolean;
     repoTasksPath?: string;
     skipEval: boolean;
     skipFetch: boolean;
@@ -152,7 +150,7 @@ export interface PlanOptions {
     taskOption?: string;
 }
 /**
- * Build a complete execution plan for the `pipeline` command.
+ * Build a complete execution plan for the `run` command.
  *
  * This is a read-only operation — it computes the plan by calling existing
  * pure functions (task expansion, model loading, cache lookup, pricing)

package/dist/pipeline/plan.js

@@ -99,7 +99,7 @@ const AVG_TOKENS = {
 // Cache prediction
 // ---------------------------------------------------------------------------
 /**
- * Build a complete execution plan for the `pipeline` command.
+ * Build a complete execution plan for the `run` command.
  *
  * This is a read-only operation — it computes the plan by calling existing
  * pure functions (task expansion, model loading, cache lookup, pricing)
@@ -322,13 +322,11 @@ export async function buildPipelinePlan(opts, rootDir) {
     // 7. Build step plan
     const steps = buildStepPlan({
         compareEnabled: opts.compareEnabled,
-        discoveryReportEnabled: opts.discoveryReportEnabled,
         dryRun: opts.dryRun,
         gapAnalysisEnabled: opts.gapAnalysisEnabled,
         graderReplications: opts.graderReplications,
         noCache: opts.noCache,
         publishEnabled: opts.publishEnabled,
-        readinessEnabled: opts.readinessEnabled,
         skipEval: opts.skipEval,
         skipFetch: opts.skipFetch,
     }, cachePrediction);
@@ -338,14 +336,12 @@ export async function buildPipelinePlan(opts, rootDir) {
     const filesRead = collectFilesRead(rootDir, opts.mode);
     const filesCreated = collectFilesCreated({
         compareEnabled: opts.compareEnabled,
-        discoveryReportEnabled: opts.discoveryReportEnabled,
         gapAnalysisEnabled: opts.gapAnalysisEnabled,
         publishEnabled: opts.publishEnabled,
-        readinessEnabled: opts.readinessEnabled,
     });
     return {
         cacheStatus: cachePrediction.predictions,
-        command: "pipeline",
+        command: "run",
         comparison,
         costEstimate,
         debug: debugPlan,
@@ -486,7 +482,7 @@ function buildStepPlan(opts, cachePrediction) {
         estimatedSavedMs: fetchStatus === "hit" ? cachePrediction.estimatedSavedMs : undefined,
         name: "Fetch documentation",
         reason: fetchStatus === "skipped"
-            ? "--skip-fetch: reuse cached contexts"
+            ? "--no-fetch: reuse cached contexts"
             : fetchStatus === "hit"
                 ? "CACHED (inputs unchanged)"
                 : "Fetch from Sanity Content Lake",
@@ -508,7 +504,7 @@ function buildStepPlan(opts, cachePrediction) {
         cacheStatus: evalStatus,
         name: "Run evaluation",
         reason: opts.skipEval
-            ? "--skip-eval: reuse existing results"
+            ? "--no-eval: reuse existing results"
             : evalStatus === "hit"
                 ? "CACHED (inputs unchanged)"
                 : "Execute Promptfoo evaluation against all models",
@@ -567,24 +563,6 @@ function buildStepPlan(opts, cachePrediction) {
             willRun: true,
         });
     }
-    // Step 6: Readiness report (optional)
-    if (opts.readinessEnabled) {
-        steps.push({
-            cacheStatus: "miss",
-            name: "Readiness report",
-            reason: "Generate launch readiness checklist",
-            willRun: true,
-        });
-    }
-    // Step 6c: Discovery report (optional)
-    if (opts.discoveryReportEnabled) {
-        steps.push({
-            cacheStatus: "miss",
-            name: "Discovery report",
-            reason: "Analyze agent discoverability from retrieval metrics",
-            willRun: true,
-        });
-    }
     return steps;
 }
 function collectFilesCreated(opts) {
@@ -601,12 +579,6 @@ function collectFilesCreated(opts) {
         files.push("results/latest/failure-modes.json");
         files.push("results/latest/gap-analysis.json");
     }
-    if (opts.readinessEnabled) {
-        files.push("results/latest/readiness-report.md");
-    }
-    if (opts.discoveryReportEnabled) {
-        files.push("results/latest/discovery-report.md");
-    }
     return files.sort();
 }
 // ---------------------------------------------------------------------------

package/dist/pipeline/run-context.d.ts

@@ -85,7 +85,7 @@ export interface RunContextInput {
 export declare function buildRunContext(input: RunContextInput): RunContext;
 /**
  * Resolve `classification` from `AILF_CLASSIFICATION`, validated against
- * the closed enum. Defaults to `"ad-hoc"` so unannotated runs never leak
+ * the closed enum. Defaults to `"adhoc"` so unannotated runs never leak
  * into the canonical `"official"` series.
  */
 export declare function detectClassification(log: Logger): RunClassification;

package/dist/pipeline/run-context.js

@@ -169,17 +169,17 @@ function detectTrigger() {
 // ---------------------------------------------------------------------------
 /**
  * Resolve `classification` from `AILF_CLASSIFICATION`, validated against
- * the closed enum. Defaults to `"ad-hoc"` so unannotated runs never leak
+ * the closed enum. Defaults to `"adhoc"` so unannotated runs never leak
  * into the canonical `"official"` series.
  */
 export function detectClassification(log) {
     const raw = process.env.AILF_CLASSIFICATION?.trim();
     if (!raw)
-        return "ad-hoc";
+        return "adhoc";
     if (isRunClassification(raw))
         return raw;
-    log.warn(`AILF_CLASSIFICATION="${raw}" is not a recognized value; defaulting to "ad-hoc"`);
-    return "ad-hoc";
+    log.warn(`AILF_CLASSIFICATION="${raw}" is not a recognized value; defaulting to "adhoc"`);
+    return "adhoc";
 }
 /**
  * Resolve `owner` from `AILF_OWNER_TEAM` (+ optional

package/dist/pipeline/validate.d.ts

@@ -63,6 +63,6 @@ export declare function validateTaskFiles(rootDir: string): ValidationIssue[];
  *
  * Returns warnings (not errors) if the file is missing — thresholds are
  * optional and don't block evaluation. They only activate when
- * `--readiness` or severity-aware sink routing is used.
+ * `ailf report readiness` or severity-aware sink routing is used.
  */
 export declare function validateThresholdsYaml(rootDir: string): ValidationIssue[];

package/dist/pipeline/validate.js

@@ -272,7 +272,7 @@ export function validateTaskFiles(rootDir) {
  *
  * Returns warnings (not errors) if the file is missing — thresholds are
  * optional and don't block evaluation. They only activate when
- * `--readiness` or severity-aware sink routing is used.
+ * `ailf report readiness` or severity-aware sink routing is used.
  */
 export function validateThresholdsYaml(rootDir) {
     const source = "validateThresholdsYaml";