> For the complete documentation index, see [llms.txt](https://ai-os-and-trend-finder.gitbook.io/ai-os-and-trend-finder-docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://ai-os-and-trend-finder.gitbook.io/ai-os-and-trend-finder-docs/.spec_system/archive/sessions/phase27-session08-dated-predictions-and-story-log/implementation-notes.md). # Implementation Notes **Session ID**: `phase27-session08-dated-predictions-and-story-log` **Started**: 2026-06-13 02:30 **Last Updated**: 2026-06-13 03:16 *** ## Session Progress | Metric | Value | | ------------------- | --------- | | Tasks Completed | 22 / 22 | | Estimated Remaining | 0 minutes | | Blockers | 0 | *** ## Task Log ### 2026-06-13 - Session Start **Environment verified**: * [x] Prerequisites confirmed * [x] Tools available * [x] Directory structure ready * [x] Bun 1.3.14 available *** ### Task T001 - Verify analyzer state and prerequisites **Started**: 2026-06-13 02:29 **Completed**: 2026-06-13 02:30 **Duration**: 1 minute **Notes**: * Ran `.spec_system/scripts/analyze-project.sh --json`; current session is `phase27-session08-dated-predictions-and-story-log`. * Confirmed Phase 27 Session 01 and Session 06 are listed in completed sessions by analyzer state. * Read the active spec and confirmed dated prediction plus Story Log scope is additive and browser-safe. * Read `.spec_system/SECURITY-COMPLIANCE.md`; current posture is CLEAN with 0 open findings. * Ran `.spec_system/scripts/check-prereqs.sh --json --env` and `--tools "bun"`; both passed. **Files Changed**: * `.spec_system/specs/phase27-session08-dated-predictions-and-story-log/implementation-notes.md` - Created implementation log and recorded setup verification. *** ### Task T002 - Inspect prediction, retro, collector, schema, view-model, fixture, and test baselines **Started**: 2026-06-13 02:30 **Completed**: 2026-06-13 02:31 **Duration**: 1 minute **Notes**: * Inspected prediction schema and browser summary contracts in `src/extensions/trend-finder/schema.ts`. * Inspected prediction archive, fallback, prompt, AI output validation, and cap helpers in `scripts/lib/ai-runtime/predictions.ts`. * Inspected retro archive, calibration, pending/stale logic, prompt, and AI output validation in `scripts/lib/ai-runtime/retros.ts`. * Inspected collector prediction-retro pipeline, candidate wiring, archive reads/writes, trace decisions, and snapshot publish path in `scripts/extensions/trend-finder/collector.ts`. * Inspected Watchlist and Brief prediction-retro render paths plus existing view-model projections. * Inspected fixture, schema, collector, prediction, retro, view-model, and component test patterns. * Read `docs/adr/0001-extension-platform-foundation.md`; changes must stay inside static typed extension boundaries and browser-safe extension data. **Files Changed**: * `.spec_system/specs/phase27-session08-dated-predictions-and-story-log/implementation-notes.md` - Recorded baseline inspection. *** ### Task T003 - Add dated prediction and Story Log schemas **Started**: 2026-06-13 02:31 **Completed**: 2026-06-13 02:34 **Duration**: 3 minutes **Notes**: * Extended forecast horizons with deterministic dated buckets: `30-day`, `60-day`, and `90-day`. * Added optional `targetDate`, `targetDateLabel`, `lifecycleAtFiling`, and `targetLifecycle` fields to prediction records. * Added retro `dueState` plus optional dated/lifecycle fields for target-aware display and grading. * Added bounded `PredictionStoryLogRowSchema` and a 50-row `storyLog` branch under `PredictionRetroSummarySchema`. * Kept legacy defaults additive through existing schema error mapping and `DEFAULT_PREDICTION_RETRO_SUMMARY`. * Ran `bun run test src/lib/__tests__/trend-finder-schema.test.ts`: 91 tests passed. **Files Changed**: * `src/extensions/trend-finder/schema.ts` - Added dated prediction fields, due-state enum, Story Log row schema, browser cap, and summary defaults. **BQC Fixes**: * Contract alignment: preserved legacy prediction and retro payload parsing with additive defaults and focused schema coverage. *** ### Task T004 - Extend prediction prompt validation and sanitization for dated fields **Started**: 2026-06-13 02:34 **Completed**: 2026-06-13 02:36 **Duration**: 2 minutes **Notes**: * Added target-date and lifecycle guidance to prediction system/task prompts and output schema hints. * Added optional lifecycle stage input support on prediction topic candidates. * Added explicit validation for forecast horizon enum values, UTC target dates, and lifecycle stage enums. * Preserved fallback behavior for invalid AI rows while sanitizing accepted dated display fields. * Ran `bun run test scripts/lib/ai-runtime/__tests__/predictions.test.ts`: 19 tests passed. **Files Changed**: * `scripts/lib/ai-runtime/predictions.ts` - Extended AI candidate validation, prompt context, output schema hint, and accepted prediction record fields. **BQC Fixes**: * Trust boundary enforcement: invalid target dates, unknown lifecycle stages, and invalid horizons now reject AI rows before browser/archive persistence. * Contract alignment: horizon labels now use an exhaustive enum switch. *** ### Task T005 - Implement deterministic dated fallback buckets **Started**: 2026-06-13 02:36 **Completed**: 2026-06-13 02:38 **Duration**: 2 minutes **Notes**: * Added deterministic 30/60/90-day fallback horizon selection from topic lifecycle, movement, and expected direction. * Added UTC target-date arithmetic based only on `generatedAt`; invalid generated dates fall back to a fixed epoch base rather than wall-clock time. * Added fallback lifecycle transition targets for rising predictions while preserving current stage for stable/cooling predictions. * Updated the stale prediction test expectation from `next-run` to the new dated 30-day fallback contract. * Ran `bun run test scripts/lib/ai-runtime/__tests__/predictions.test.ts`: 19 tests passed. **Files Changed**: * `scripts/lib/ai-runtime/predictions.ts` - Added deterministic horizon bucket, target date, and target lifecycle fallback helpers. * `scripts/lib/ai-runtime/__tests__/predictions.test.ts` - Updated deterministic fallback expectation for dated prediction output. **BQC Fixes**: * State freshness on re-entry: fallback target dates are derived from the input run timestamp every time and never from process wall-clock state. * Contract alignment: deterministic fallback now writes the same dated fields as accepted AI output. *** ### Task T006 - Add target-date eligibility and stale-pending behavior **Started**: 2026-06-13 02:38 **Completed**: 2026-06-13 02:42 **Duration**: 4 minutes **Notes**: * Added UTC date-only due-state evaluation for prediction target dates. * Preserved dated fields on sanitized retro records. * Converted not-yet-due dated predictions into pending retro rows before AI/runtime grading. * Excluded not-yet-due predictions from the retro model prompt and merged their pending rows back into results. * Prevented not-yet-due pending rows from incrementing stale-pending archive counts. * Ran `bun run test scripts/lib/ai-runtime/__tests__/retros.test.ts`: 26 tests passed. **Files Changed**: * `scripts/lib/ai-runtime/retros.ts` - Added target-date due-state helpers, not-yet-due pending behavior, prompt context, and stale-pending count guard. **BQC Fixes**: * Trust boundary enforcement: dated grading is enforced before AI prompt construction and again during candidate conversion. * Failure path completeness: runtime/parse fallbacks preserve not-yet-due pending rows instead of dropping them. * Contract alignment: retro records now carry the dated fields needed by browser display. *** ### Task T007 - Extend calibration helpers for dated outcomes **Started**: 2026-06-13 02:42 **Completed**: 2026-06-13 02:43 **Duration**: 1 minute **Notes**: * Added calibration eligibility filtering for retro rows. * Kept resolved dated outcomes in hit-rate and Brier-style math. * Excluded `not-yet-due` pending rows from calibration total and pending outcome counts. * Added explicit unavailable copy for archives that contain only not-yet-due targets. * Ran `bun run test scripts/lib/ai-runtime/__tests__/retros.test.ts`: 26 tests passed. **Files Changed**: * `scripts/lib/ai-runtime/retros.ts` - Updated calibration summary math to exclude not-yet-due pending rows. **BQC Fixes**: * Contract alignment: calibration counters now match dated due-state semantics and avoid unresolved target inflation. *** ### Task T008 - Define Story Log view-model types and filters **Started**: 2026-06-13 02:43 **Completed**: 2026-06-13 02:45 **Duration**: 2 minutes **Notes**: * Added Story Log row and summary view-model types. * Added verdict and due-state filter option metadata with validated filter normalization. * Added due-state labels and tones. * Added deterministic Story Log row ordering by due state, target date, generated date, topic name, prediction ID, and retro ID. * Added stable 50-row view-model bound matching the schema cap. * Ran `bun run test src/extensions/trend-finder/__tests__/view-model.test.ts`: 23 tests passed. **Files Changed**: * `src/extensions/trend-finder/view-model.ts` - Added Story Log projection, filters, labels, sorting, and summary integration. **BQC Fixes**: * State freshness on re-entry: filter normalization clamps invalid filter state to `all`. * Contract alignment: Story Log sorting and row bounds are deterministic before UI rendering. *** ### Task T009 - Pass topic lifecycle into prediction candidates **Started**: 2026-06-13 02:45 **Completed**: 2026-06-13 02:46 **Duration**: 1 minute **Notes**: * Passed `topic.lifecycle.stage` into prediction candidates with an explicit `unknown` fallback. * This gives AI prompt context and deterministic fallback bucket selection the same lifecycle source. * Ran `bun run test scripts/extensions/trend-finder/__tests__/collector.test.ts`: 21 tests passed. **Files Changed**: * `scripts/extensions/trend-finder/collector.ts` - Added lifecycle stage to prediction topic candidate construction. **BQC Fixes**: * Contract alignment: collector candidate fields now match the prediction runtime candidate schema. *** ### Task T010 - Build capped Story Log projection **Started**: 2026-06-13 02:46 **Completed**: 2026-06-13 02:48 **Duration**: 2 minutes **Notes**: * Added collector Story Log projection from sanitized prediction and retro records. * Dedupe logic keeps retro rows authoritative and adds prediction-only rows only when no retro exists for the prediction. * Added deterministic Story Log ordering and a 50-row browser cap. * Preserved recent prediction and retro summary caps separately from the new Story Log cap. * Ran `bun run test scripts/extensions/trend-finder/__tests__/collector.test.ts`: 21 tests passed. **Files Changed**: * `scripts/extensions/trend-finder/collector.ts` - Added Story Log projection helpers and attached `storyLog` to `predictionRetroSummary`. **BQC Fixes**: * Contract alignment: Story Log rows are built from sanitized runtime records and match the schema branch. * Error information boundaries: projection never includes archive paths, prompts, provider responses, or raw archive payloads. *** ### Task T011 - Emit sanitized prediction history trace counters **Started**: 2026-06-13 02:48 **Completed**: 2026-06-13 02:51 **Duration**: 3 minutes **Notes**: * Added a typed `prediction-history` Engine Replay artifact ID and client labels/icons. * Emitted a sanitized prediction-history decision with aggregate Story Log counts only. * Trace reason includes row, resolved, due, not-yet-due, and pending counts; it excludes raw rows, archive paths, prompts, and provider responses. * Ran `bun run test scripts/extensions/trend-finder/__tests__/collector.test.ts`: 21 tests passed. * Ran `bun run test src/lib/__tests__/trend-finder-schema.test.ts src/lib/__tests__/trend-finder-engine-replay.test.tsx`: 110 tests passed. **Files Changed**: * `scripts/extensions/trend-finder/collector.ts` - Added sanitized prediction-history trace decision and counters. * `src/extensions/trend-finder/engine-trace.ts` - Added `prediction-history` artifact ID. * `src/extensions/trend-finder/engine-replay-model.ts` - Added artifact title and description. * `src/extensions/trend-finder/components/engine-artifact-rail.tsx` - Added artifact icon mapping. **BQC Fixes**: * Error information boundaries: Engine Replay receives aggregate counters only and no raw prediction or retro records. * Contract alignment: new trace artifact ID is registered in schema and rendering model. *** ### Task T012 - Create Story Log panel **Started**: 2026-06-13 02:51 **Completed**: 2026-06-13 02:53 **Duration**: 2 minutes **Notes**: * Created a dense Story Log panel with verdict and due-state segmented filters. * Added explicit empty and filtered-empty states. * Added reset behavior and automatic filter reset when the row signature changes. * Added accessible labels, pressed state, section heading, and reset control. * Deferred component tests until T021, when the component is imported and exercised with fixtures. **Files Changed**: * `src/extensions/trend-finder/components/story-log-panel.tsx` - New filterable Story Log panel. **BQC Fixes**: * Accessibility and platform compliance: filter controls use button semantics, `aria-pressed`, grouped labels, and reset labeling. * State freshness on re-entry: filters reset when the Story Log row set changes. *** ### Task T013 - Wire Story Log panel into Watchlist **Started**: 2026-06-13 02:53 **Completed**: 2026-06-13 02:54 **Duration**: 1 minute **Notes**: * Added the Story Log panel to Watchlist as a global prediction-history panel above generated rows. * Added a Watchlist visibility-menu entry for the Story Log panel when prediction archives are present. * Preserved existing per-topic prediction/retro sections and missing-topic row rendering. * Ran `bun run test src/extensions/trend-finder/__tests__/visibility-views.test.tsx`: 10 tests passed. * Ran `bun run test src/extensions/trend-finder/__tests__/view-model.test.ts`: 23 tests passed. **Files Changed**: * `src/extensions/trend-finder/views/watchlist-view.tsx` - Imported and rendered `StoryLogPanel`, added visibility entry, and preserved mobile-safe vertical stacking. **BQC Fixes**: * State freshness on re-entry: the Watchlist panel delegates filter reset/revalidation to the Story Log component on data changes. * Accessibility and platform compliance: the panel remains inside the existing visibility controls and focusable control model. *** ### Task T014 - Add compact prediction-history summary to Brief **Started**: 2026-06-13 02:54 **Completed**: 2026-06-13 02:55 **Duration**: 1 minute **Notes**: * Added a compact Prediction history subsection inside the existing Brief prediction/retro panel. * Shows up to three Story Log rows with verdict, due state, target date label, provenance, topic, and recent lesson/rationale. * Added explicit empty copy when no Story Log rows are available. * Ran `bun run test src/extensions/trend-finder/__tests__/visibility-views.test.tsx`: 10 tests passed. * Ran `bun run test src/extensions/trend-finder/__tests__/view-model.test.ts`: 23 tests passed. **Files Changed**: * `src/extensions/trend-finder/views/brief-view.tsx` - Added compact Story Log summary to the Brief prediction/retro section. **BQC Fixes**: * Accessibility and platform compliance: Story Log summary rows preserve row ARIA labels in Brief. * Contract alignment: Brief renders the same view-model Story Log projection as Watchlist. *** ### Task T015 - Add dated prediction and Story Log fixture states **Started**: 2026-06-13 02:55 **Completed**: 2026-06-13 02:57 **Duration**: 2 minutes **Notes**: * Added target date, lifecycle-at-filing, target lifecycle, and dated horizon fields to fixture predictions and retros. * Added Story Log rows covering not-yet-due pending, due stale pending, and resolved hit states. * Added exported Story Log fixture rows plus legacy-omitted and malformed Story Log payload variants. * Ran `bun run test src/lib/__tests__/trend-finder-schema.test.ts`: 91 tests passed. * Ran `bun run test src/extensions/trend-finder/__tests__/view-model.test.ts src/extensions/trend-finder/__tests__/visibility-views.test.tsx`: 33 tests passed. **Files Changed**: * `src/extensions/trend-finder/fixtures.ts` - Added dated prediction, retro, Story Log, malformed, and legacy fixture states. **BQC Fixes**: * Contract alignment: browser fixtures now exercise the new additive Story Log and dated prediction fields. *** ### Task T016 - Update committed example payload **Started**: 2026-06-13 02:57 **Completed**: 2026-06-13 02:58 **Duration**: 1 minute **Notes**: * Added a compact additive `predictionRetroSummary` branch to the Trend Finder example extension data. * Included one dated prediction, one resolved retro, and one Story Log row. * Kept the example browser-safe: no raw archive paths, prompts, provider responses, cache paths, or private metadata. * Ran `jq empty src/data/live-data.example.json`: passed. * Ran `bun run test src/lib/__tests__/trend-finder-schema.test.ts`: 91 tests passed. **Files Changed**: * `src/data/live-data.example.json` - Added compact dated prediction and Story Log example data. **BQC Fixes**: * Contract alignment: committed fallback data now includes the additive dated prediction and Story Log branches. * Error information boundaries: example payload contains only browser-safe summarized fields. *** ### Task T017 - Verify archive read/write atomicity and warning compensation **Started**: 2026-06-13 02:58 **Completed**: 2026-06-13 02:59 **Duration**: 1 minute **Notes**: * Verified prediction and retro archive tests still cover atomic writes, latest pointer updates, malformed reads, missing reads, and pending listing. * Verified collector tests still cover write failure warning compensation and browser-safe summary emission. * No additional code changes were needed beyond the dated field support already added to prediction and retro record sanitizers. * Ran `bun run test scripts/lib/ai-runtime/__tests__/predictions.test.ts scripts/lib/ai-runtime/__tests__/retros.test.ts`: 45 tests passed. * Ran `bun run test scripts/extensions/trend-finder/__tests__/collector.test.ts`: 21 tests passed. **Files Changed**: * No new files changed for this verification task. **BQC Fixes**: * Failure path completeness: focused collector coverage confirms archive write failures remain warnings, not silent data loss. *** ### Task T018 - Extend prediction runtime tests **Started**: 2026-06-13 02:59 **Completed**: 2026-06-13 03:00 **Duration**: 1 minute **Notes**: * Added tests for deterministic 30/60/90-day fallback buckets from lifecycle and movement. * Extended valid AI response coverage for target date, target label, lifecycle-at-filing, and target lifecycle. * Added invalid target-date and unknown lifecycle-stage rejection tests. * Added legacy archive parsing coverage for records without dated fields. * Existing unknown evidence, sanitization, archive atomicity, and browser cap tests remain passing. * Ran `bun run test scripts/lib/ai-runtime/__tests__/predictions.test.ts`: 23 tests passed. **Files Changed**: * `scripts/lib/ai-runtime/__tests__/predictions.test.ts` - Added dated prediction runtime and archive compatibility tests. **BQC Fixes**: * Trust boundary enforcement: tests now cover invalid target dates and unknown lifecycle stages at AI output boundaries. *** ### Task T019 - Extend retro runtime tests **Started**: 2026-06-13 03:00 **Completed**: 2026-06-13 03:01 **Duration**: 1 minute **Notes**: * Added before-target not-yet-due pending coverage. * Added on-target deterministic grading coverage. * Added after-target stale-pending coverage for disappeared topics. * Added calibration coverage that excludes not-yet-due pending rows from totals and outcome counts. * Added runtime enforcement coverage proving not-yet-due predictions are not sent to the AI retro evaluator. * Existing disappeared-topic, backtest mode preservation, and calibration math tests remain passing. * Ran `bun run test scripts/lib/ai-runtime/__tests__/retros.test.ts`: 31 tests passed. **Files Changed**: * `scripts/lib/ai-runtime/__tests__/retros.test.ts` - Added dated retro grading-window and calibration tests. **BQC Fixes**: * Trust boundary enforcement: tests prove not-yet-due rows are locally enforced before AI runtime calls. *** ### Task T020 - Extend collector and schema tests **Started**: 2026-06-13 03:01 **Completed**: 2026-06-13 03:06 **Duration**: 5 minutes **Notes**: * Added schema coverage for legacy payloads that omit `storyLog`, populated Story Log rows, malformed Story Log rows, and the 50-row browser cap. * Extended collector coverage to assert lifecycle stage wiring into prediction candidates. * Added collector assertions for capped Story Log projection rows and sanitized `prediction-history` trace artifact counters. * Confirmed browser output avoids raw archive/cache path leakage in the prediction history trace. * Ran `bun run test src/lib/__tests__/trend-finder-schema.test.ts scripts/extensions/trend-finder/__tests__/collector.test.ts`: 116 tests passed. **Files Changed**: * `src/lib/__tests__/trend-finder-schema.test.ts` - Added Story Log schema compatibility, malformed payload, and cap tests. * `scripts/extensions/trend-finder/__tests__/collector.test.ts` - Added lifecycle candidate, Story Log projection, and sanitized trace counter assertions. **BQC Fixes**: * Contract alignment: tests now cover additive Story Log defaults and browser caps. * Error information boundaries: collector tests prove prediction history trace output uses aggregate counts, not private paths. *** ### Task T021 - Add Story Log view-model and component tests **Started**: 2026-06-13 03:06 **Completed**: 2026-06-13 03:08 **Duration**: 2 minutes **Notes**: * Added view-model coverage for Story Log total counts, due-state ordering, verdict filter counts, due-state filter counts, invalid filter normalization, and filtered pending/not-yet-due rows. * Added Story Log panel tests for filter ARIA labels, row ARIA labels, mobile-safe row text, empty state copy, reset button behavior, and reset on row-signature changes. * Ran `bun run test src/extensions/trend-finder/__tests__/view-model.test.ts src/extensions/trend-finder/components/__tests__/story-log-panel.test.tsx`: 27 tests passed. **Files Changed**: * `src/extensions/trend-finder/__tests__/view-model.test.ts` - Added Story Log filter/count/order assertions. * `src/extensions/trend-finder/components/__tests__/story-log-panel.test.tsx` - Added focused component coverage for Story Log interactions and empty states. **BQC Fixes**: * State freshness on re-entry: component tests prove active filters reset when the Story Log row signature changes. * Accessibility and platform compliance: tests cover ARIA labels on filters and rows plus compact row text. *** ### Task T022 - Run focused validation and record implementation results **Started**: 2026-06-13 03:08 **Completed**: 2026-06-13 03:16 **Duration**: 8 minutes **Notes**: * Ran focused validation across prediction, retro, collector, schema, view-model, visibility, Story Log component, and Engine Replay tests. * Verified `bun run typecheck` passes. * Verified `bun run typecheck:scripts` passes. * Verified `bunx prettier --check` passes for touched TypeScript and TSX files after formatting. * Verified `jq empty src/data/live-data.example.json` passes. * Verified changed/new files contain ASCII-only characters. * Verified `src/data/live-data.example.json` is 46,117 bytes, under the shared 1 MB payload limit. * Verified `bun run budget:check` passes with 0 violations. * Verified full `bun run test` passes: 278 test files and 3424 tests. * Ran `bun run test scripts/lib/ai-runtime/__tests__/predictions.test.ts scripts/lib/ai-runtime/__tests__/retros.test.ts scripts/extensions/trend-finder/__tests__/collector.test.ts src/lib/__tests__/trend-finder-schema.test.ts src/extensions/trend-finder/__tests__/view-model.test.ts src/extensions/trend-finder/__tests__/visibility-views.test.tsx src/extensions/trend-finder/components/__tests__/story-log-panel.test.tsx src/lib/__tests__/trend-finder-engine-replay.test.tsx`: 226 tests passed. **Files Changed**: * `scripts/lib/ai-runtime/predictions.ts` - Made fallback lifecycle horizon switch exhaustiveness visible to TypeScript. * `scripts/lib/ai-runtime/backtests.ts` - Preserved the additive `storyLog` summary field in backtest snapshot summaries. * `scripts/trend-finder-replay-backhistory.ts` - Preserved the additive `storyLog` summary field in replay snapshot summaries. * `scripts/lib/ai-runtime/__tests__/predictions.test.ts` - Added explicit due-state fields to direct retro test records. * `scripts/lib/ai-runtime/__tests__/retros.test.ts` - Derived and asserted due-state fields in retro test helpers and direct records. * `scripts/extensions/trend-finder/__tests__/collector.test.ts` - Added explicit due-state field to the mocked retro result. * `scripts/extensions/trend-finder/__tests__/historical-backhistory-flow.test.ts` - Updated the full-flow expectation for dated not-yet-due replay predictions. * `.spec_system/specs/phase27-session08-dated-predictions-and-story-log/implementation-notes.md` - Recorded final validation. * `.spec_system/specs/phase27-session08-dated-predictions-and-story-log/tasks.md` - Marked T022 and completion checklist complete. **BQC Fixes**: * Contract alignment: final type checks caught and fixed missing `storyLog` fields in script-side summary builders and missing required due-state fields in direct retro records. * Contract alignment: focused tests, type checks, schema checks, and UI checks now agree on the dated prediction and Story Log contracts. ## Final Implementation Summary * Added additive dated prediction fields with deterministic 30/60/90-day fallback buckets, target dates, lifecycle-at-filing, target lifecycle, and AI validation for dated/lifecycle output. * Added target-date-aware retro grading with not-yet-due pending behavior, due/stale pending semantics, and calibration that excludes not-yet-due rows. * Added a capped browser-safe Story Log projection with sanitized prediction/retro history, aggregate Engine Replay counters, and no raw archive/cache path exposure. * Rendered Story Log filters in Watchlist and compact prediction history in Brief using validated view-model filters and reset-on-data-change behavior. * Added fixture, example payload, schema, collector, runtime, view-model, and component coverage for legacy defaults, malformed rows, caps, filters, ordering, accessibility, and replay integration. * All 22 tasks are complete and the session is ready for the validate workflow step. --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter: ``` GET https://ai-os-and-trend-finder.gitbook.io/ai-os-and-trend-finder-docs/.spec_system/archive/sessions/phase27-session08-dated-predictions-and-story-log/implementation-notes.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.