bitfab-cli 0.2.314 → 0.2.315

package/dist/index.js

@@ -23934,7 +23934,7 @@ var getTraces = {
 var getTraceLabels = {
   name: "get_trace_labels",
   title: "Get Trace Labels",
-  description: `Read just the labels for one or more traces by ID: returns each trace's pass/fail verdict, its annotation (the reviewer's reasoning), and whether the label is human-validated (a human authored or approved it). No span content, inputs, or outputs are loaded, so the response is small and one call accepts up to ${GET_TRACE_LABELS_MAX_IDS} IDs. Use this to load a whole dataset's verdicts into context in a single call (e.g. when building or confirming a dataset). When you need a trace's actual inputs/outputs/spans, use get_traces instead. This returns verdicts on runs that already happened; for what a trace SHOULD do on its next replay, use get_trace_assertions.`,
+  description: `Read just the labels for one or more traces by ID: returns each trace's pass/fail verdict, its annotation (the reviewer's reasoning), and whether the label is human-validated (a human authored or approved it). Per-assertion verdicts follow on indented lines, one per assertion, keyed by the assertionId the write used, each with its own verdict, annotation, confidence, and author, so a verdict written per assertion reads back per assertion rather than as a tally. No span content, inputs, or outputs are loaded, so the response is small and one call accepts up to ${GET_TRACE_LABELS_MAX_IDS} IDs. Use this to load a whole dataset's verdicts into context in a single call (e.g. when building or confirming a dataset). When you need a trace's actual inputs/outputs/spans, use get_traces instead. This returns verdicts on runs that already happened; for what a trace SHOULD do on its next replay, use get_trace_assertions.`,
   inputSchema: {
     traceIds: external_exports.preprocess(parseJsonString, external_exports.array(external_exports.uuid()).min(1).max(GET_TRACE_LABELS_MAX_IDS)).describe(`Trace IDs to read labels for (1-${GET_TRACE_LABELS_MAX_IDS})`)
   }
@@ -23953,7 +23953,7 @@ var getSpanField = {
 var saveAgentLabels = {
   name: "save_agent_labels",
   title: "Save Agent Labels",
-  description: "Set, skip, or archive the agent's pass/fail verdict on one or more traces (`labelSource=\"agent\"`). Use this AFTER you have read the traces with get_traces and decided yourself whether each one looks like a pass, a fail, or genuinely cannot be judged. To set a verdict, pass `label` (true=PASS, false=FAIL) and `annotation` (your reasoning, shown to the human reviewer in the labeling UI). Optionally pass `confidence` (`VeryLow|Low|Medium|High|VeryHigh`) to record how confident you are - surfaced in the labeling UI so reviewers can prioritize low-confidence verdicts. To explicitly skip a trace you cannot decide on (instead of leaving it unlabeled), pass `skip: true` and omit label/annotation - this records an explicit skip so coverage checks know you intentionally did not verdict it. To clear a previously-set agent verdict (e.g., you changed your mind or labeled the wrong trace), pass `archive: true` and omit label/annotation. New verdicts start unapproved (`approvedAt=null`); once a human approves one, it joins the validated dataset (queryable via `search_traces` with `validated: true`). Archiving is non-destructive - the row is hidden from all reads but kept for audit, and you can immediately re-label the trace from scratch. For replay results, key each label by the replay item's `originalTraceId` (the original trace it was replayed from; `sourceTraceId` is accepted as a deprecated alias) plus the top-level `testRunId` instead of a `traceId`: the server resolves it to the replay trace via lineage, so you never need a server-generated replay trace id. When the experiment ran each trace more than once, also pass the item's `attempt` so each attempt gets its own verdict. Returns an agent-readable summary with one parseable effective label line per updated trace, keyed by the id you supplied (`originalTraceId` for replay verdicts, otherwise `traceId`), so command callers can verify persistence. Before judging a replay, call get_trace_assertions on the original trace ids: an expectation says what the user asked this specific case to do, so the verdict is measured against that rather than a guess. When an expectation's target cannot be found on the trace you are judging, the check never ran, so pass `skip: true` for that trace rather than recording a FAIL. Pass `assertionId` to score ONE assertion on the trace (the `[ID: ...]` get_trace_assertions prints), one entry per assertion, all in the same call. Omit it for the trace's single whole-trace verdict. get_trace_labels reports the resulting passed and failed counts per trace.",
+  description: "Set, skip, or archive the agent's pass/fail verdict on one or more traces (`labelSource=\"agent\"`). Use this AFTER you have read the traces with get_traces and decided yourself whether each one looks like a pass, a fail, or genuinely cannot be judged. To set a verdict, pass `label` (true=PASS, false=FAIL) and `annotation` (your reasoning, shown to the human reviewer in the labeling UI). Optionally pass `confidence` (`VeryLow|Low|Medium|High|VeryHigh`) to record how confident you are - surfaced in the labeling UI so reviewers can prioritize low-confidence verdicts. To explicitly skip a trace you cannot decide on (instead of leaving it unlabeled), pass `skip: true` and omit label/annotation - this records an explicit skip so coverage checks know you intentionally did not verdict it. To clear a previously-set agent verdict (e.g., you changed your mind or labeled the wrong trace), pass `archive: true` and omit label/annotation. New verdicts start unapproved (`approvedAt=null`); once a human approves one, it joins the validated dataset (queryable via `search_traces` with `validated: true`). Archiving is non-destructive - the row is hidden from all reads but kept for audit, and you can immediately re-label the trace from scratch. For replay results, key each label by the replay item's `originalTraceId` (the original trace it was replayed from; `sourceTraceId` is accepted as a deprecated alias) plus the top-level `testRunId` instead of a `traceId`: the server resolves it to the replay trace via lineage, so you never need a server-generated replay trace id. When the experiment ran each trace more than once, also pass the item's `attempt` so each attempt gets its own verdict. Returns an agent-readable summary with one parseable effective label line per updated trace, keyed by the id you supplied (`originalTraceId` for replay verdicts, otherwise `traceId`), so command callers can verify persistence. Before judging a replay, call get_trace_assertions on the original trace ids: an expectation says what the user asked this specific case to do, so the verdict is measured against that rather than a guess. When an expectation's target cannot be found on the trace you are judging, the check never ran, so pass `skip: true` for that trace rather than recording a FAIL. Pass `assertionId` to score ONE assertion on the trace (the `[ID: ...]` get_trace_assertions prints), one entry per assertion, all in the same call. Omit it for the trace's single whole-trace verdict. get_trace_labels reads these back one line per assertion, keyed by the same assertionId, so a per-assertion write is verifiable per assertion.",
   inputSchema: {
     testRunId: external_exports.uuid().optional().describe("The replay test run id. Required when any label targets a trace by originalTraceId (replay verdicts); ignored otherwise."),
     labels: external_exports.preprocess(parseJsonString, external_exports.array(external_exports.object({
@@ -23973,7 +23973,7 @@ var saveAgentLabels = {
 var saveHumanLabels = {
   name: "save_human_labels",
   title: "Save Human Labels",
-  description: "Record one or more human-authored pass/fail verdicts that are VALIDATED immediately (`labelSource=\"human\"`, no approval step). Unlike save_agent_labels (which writes agent suggestions that start unapproved), labels set here join the validated dataset the instant they are written and satisfy `search_traces` with `validated: true`. Pass `label` (true=PASS, false=FAIL), `annotation` (the reasoning behind the verdict), and optionally `confidence` (`VeryLow|Low|Medium|High|VeryHigh`). USE ONLY when a human has explicitly decided the verdict, for example saving a known production bug as a test case. This records a verdict on a run that already happened; to record what a trace should do on its next replay, use save_trace_assertions. Do NOT use this for the agent's own first-pass guesses on traces awaiting human review; use save_agent_labels for those so they go through the normal approve/edit loop. Pass `assertionId` when the human judged ONE assertion (the `[ID: ...]` get_trace_assertions prints), one entry per assertion. Omit it for the trace's single whole-trace verdict. get_trace_labels reports the resulting passed and failed counts per trace.",
+  description: "Record one or more human-authored pass/fail verdicts that are VALIDATED immediately (`labelSource=\"human\"`, no approval step). Unlike save_agent_labels (which writes agent suggestions that start unapproved), labels set here join the validated dataset the instant they are written and satisfy `search_traces` with `validated: true`. Pass `label` (true=PASS, false=FAIL), `annotation` (the reasoning behind the verdict), and optionally `confidence` (`VeryLow|Low|Medium|High|VeryHigh`). USE ONLY when a human has explicitly decided the verdict, for example saving a known production bug as a test case. This records a verdict on a run that already happened; to record what a trace should do on its next replay, use save_trace_assertions. Do NOT use this for the agent's own first-pass guesses on traces awaiting human review; use save_agent_labels for those so they go through the normal approve/edit loop. Pass `assertionId` when the human judged ONE assertion (the `[ID: ...]` get_trace_assertions prints), one entry per assertion. Omit it for the trace's single whole-trace verdict. get_trace_labels reads these back one line per assertion, keyed by the same assertionId, so a per-assertion write is verifiable per assertion.",
   inputSchema: {
     labels: external_exports.preprocess(parseJsonString, external_exports.array(external_exports.object({
       traceId: external_exports.uuid().describe("The trace ID to label"),
@@ -24019,6 +24019,15 @@ var getTraceAssertions = {
     traceIds: external_exports.preprocess(parseJsonString, external_exports.array(external_exports.uuid()).min(1).max(GET_TRACE_ASSERTIONS_MAX_IDS)).describe(`Original trace IDs to read expectations for (1-${GET_TRACE_ASSERTIONS_MAX_IDS})`)
   }
 };
+var archiveTraceAssertions = {
+  name: "archive_trace_assertions",
+  title: "Archive Trace Expectations",
+  description: `Retire expectations on a trace so later replays stop checking them. Non-destructive: the row is hidden from get_trace_assertions and from replay judging but kept for audit, and verdicts already recorded against it stay. Use it for an expectation that was wrong or superseded; to reword one instead, pass its id back to save_trace_assertions, which edits in place. Pass the ids get_trace_assertions prints as \`[ID: ...]\`. All-or-nothing: one unknown, already-archived, or wrong-trace id fails the call naming that id and archives nothing, so re-read and retry with the active ids. Up to ${MAX_ASSERTIONS_PER_REQUEST} per call.`,
+  inputSchema: {
+    traceId: external_exports.uuid().describe("The trace the expectations are attached to. This is the ORIGINAL trace they were written on, not a replay that inherited them."),
+    assertionIds: external_exports.preprocess(parseJsonString, external_exports.array(external_exports.uuid()).min(1).max(MAX_ASSERTIONS_PER_REQUEST)).describe(`Ids of the expectations to archive (1-${MAX_ASSERTIONS_PER_REQUEST}), from get_trace_assertions`)
+  }
+};
 var saveGrader = {
   name: "save_grader",
   title: "Save Grader",
@@ -24345,6 +24354,7 @@ var ALL_TOOL_CONTRACTS = [
   saveHumanLabels,
   saveTraceAssertions,
   getTraceAssertions,
+  archiveTraceAssertions,
   saveGrader,
   listGraders,
   saveDataset,
@@ -26977,8 +26987,8 @@ var semver3 = __toESM(require_semver2(), 1);
 
 // ../bitfab-plugin-lib/dist/bakedSdkVersions.js
 var BAKED_SDK_VERSIONS = {
-  typescript: "0.48.0",
-  python: "0.49.0",
+  typescript: "0.50.0",
+  python: "0.50.0",
   ruby: "0.43.1",
   go: "0.41.1"
 };

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "bitfab-cli",
-  "version": "0.2.314",
+  "version": "0.2.315",
   "description": "Install and configure the Bitfab plugin in Claude Code, Codex, or Cursor, or the Bitfab skill pack in Amp (Alpha).",
   "type": "module",
   "main": "dist/index.js",