Skip to content

xinetzone/SpecWeave / .agents/skills/dashboard-page

Dashboard Page

Unverified

Use when the user wants to create a stable local/offline ECharts dashboard with KPI panels, time controls, auditable source modals, and optional daily data snapshots produced by automations.

Dashboard Page

Overview

Generate a polished single-folder dashboard app for recurring business metrics: an offline-capable index.html, reusable ECharts panels, KPI tiles, time-range controls, data freshness indicators, source/debug modals, and a durable data snapshot folder that automations can append to every day.

This skill is for operating dashboards and data screens, not document-style reports, slide decks, card feeds, landing pages, or narrative memos. It should feel like a dense but calm internal analytics tool: scannable, filterable, and designed for repeated use.

Use the bundled template first. Keep this file lightweight: put dashboard folder rules in references/output-layout.md, dashboard structure rules in references/dashboard-structure.md, data snapshot and automation rules in references/data-and-automation.md, shared visual interaction rules in references/visual-interactions.md, reusable browser behavior in assets/dashboard_runtime.js, and static checks in scripts/validate-echarts-dashboard.mjs.

Workflow

  1. Inspect the input data or metric sources first. Confirm date columns, grain, timezone, units, metric definitions, owner, refresh cadence, late-arriving data behavior, and source freshness before building the dashboard.
  2. Read references/output-layout.md, create one folder for the dashboard, and keep all dashboard-specific files inside that folder.
  3. Copy assets/dashboard_template.py to dashboard-slug/dashboard.py, then copy assets/dashboard_runtime.js and assets/echarts.min.js into the same folder.
  4. Put reviewed source data in dashboard-slug/data/. For recurring dashboards, store daily automation outputs in dashboard-slug/data/snapshots/.
  5. Read references/dashboard-structure.md before changing dashboard panels, shell markup, or the dashboard block model.
  6. Read references/data-and-automation.md before changing snapshot schemas, ingest behavior, time filtering, or automation instructions.
  7. Read references/visual-interactions.md before changing CSS, filter behavior, panel actions, source modals, or mobile layout.
  8. Adapt the Python data layer: input/output constants, read_sources(), normalize_snapshots(), metric summarization, panel builders, and make_dashboard_payload().
  9. Make an explicit dashboard plan before writing panels. Decide the job of each view, required filters, KPI definitions, chart payloads, and freshness warnings.
  10. Compose the dashboard in build_dashboard_blocks(payload). Do not put block assembly inside build_html(payload).
  11. Adapt dashboard-specific ECharts option builders and sourceMap entries in the JavaScript section of dashboard.py.
  12. Keep the visual and interaction contract intact unless the user explicitly asks to redesign the dashboard system.
  13. Generate dashboard-slug/index.html and dashboard-slug/dashboard_data.json. Confirm index.html starts with <!-- Generated by Trae Work -->, then run node <SKILL_ROOT>/scripts/validate-echarts-dashboard.mjs dashboard-slug/index.html.

Dashboard Contract

Read references/dashboard-structure.md when planning controls, panels, charts, tables, or changing the rendered dashboard DOM.

Block Contract

The canonical block kinds are kpi, chart, table, and note. Keep block semantics separate from rendered markup: build_dashboard_blocks(payload) composes dashboard structure, render helpers emit dashboard markup, and build_html(payload) wraps the final HTML shell. See references/dashboard-structure.md for the full contract.

Template Architecture

Start from assets/dashboard_template.py. For a new dashboard, replace:

  • input and output constants near the top;
  • read_sources();
  • normalize_snapshots();
  • summarization and time-window helpers;
  • make_dashboard_payload();
  • table-builder functions when evidence tables change;
  • build_dashboard_blocks(payload);
  • dashboard-specific ECharts option builders;
  • chartFactories, sourceMap, and initial chart registrations.

Keep reusable browser behavior in assets/dashboard_runtime.js: time-range controls, chart initialization, panel menus, Data Source modal behavior, code copy actions, and resize handling. Change this file only when the shared dashboard interaction model changes. See references/visual-interactions.md before doing so.

Keep assets/echarts.min.js bundled or otherwise provide a vetted local ECharts build. The final dashboard should inline both ECharts and dashboard_runtime.js so the HTML opens locally/offline.

Automation Contract

Read references/data-and-automation.md before designing recurring data capture. The dashboard should work with:

  • manually reviewed local source files;
  • daily JSON/JSONL/CSV snapshots written by an automation;
  • regeneration through python dashboard.py after new snapshots land.

Automation should append immutable daily snapshots instead of overwriting historical data. The dashboard should expose a visible data freshness timestamp and should let users select a date range from available snapshot dates.

Output Standard

The default deliverables are:

  • one single-file local HTML dashboard;
  • one JSON dashboard payload for review/debugging;
  • one dashboard generator script;
  • optional data/snapshots/ files used by automations.

All files should live inside the dashboard's own folder, not in the workspace root. In the final response, include the dashboard index.html path with a computer:// prefix so Trae can render it.

Assets

  • assets/dashboard_template.py: executable reference generator for an operating dashboard.
  • assets/dashboard_runtime.js: shared browser runtime for time controls, chart controls, Data Source modal, copy actions, and resize.
  • assets/echarts.min.js: bundled ECharts runtime.
  • references/output-layout.md: one-dashboard-one-folder deliverable layout and naming rules.
  • references/dashboard-structure.md: shell, filter model, panel model, and template layer contract.
  • references/data-and-automation.md: snapshot schemas, recurring ingest rules, time filtering, and automation handoff.
  • references/visual-interactions.md: styling, controls, panel actions, and modal behavior.
  • scripts/validate-echarts-dashboard.mjs: static validator for required dashboard structure and interactions.
  • agents/openai.yaml: optional model/runtime configuration for this skill.

Final Checks

Before responding to the user:

  • Run the generator and confirm the HTML and JSON payload were written.
  • Confirm the dashboard output lives in its own folder, with index.html as the main dashboard file.
  • Confirm index.html starts with <!-- Generated by Trae Work --> on line 1.
  • Run node <SKILL_ROOT>/scripts/validate-echarts-dashboard.mjs dashboard-slug/index.html.
  • Confirm the final response includes the dashboard index.html path with a computer:// prefix.
  • Confirm visible output uses dashboard structure, not document-report, card-feed, slide, or landing-page abstractions.
  • Search for placeholder text, stale sample titles, stale dates, and visible debug labels.
  • Confirm all time controls map to available payload ranges.
  • Confirm business date formatting preserves local YYYY-MM-DD dates and does not use UTC conversion such as toISOString().slice(0, 10).
  • Confirm every chart container has a registered chart factory and initial chart type.
  • Confirm panel menus include working View Data Source actions, and chart panels with alternates include Edit.
  • Confirm Data Source modal shows the panel transform snippet and analysis logic only, without exposing HTML/CSS/JS rendering or dashboard-shell generation code.
  • Confirm charts, KPI tiles, and tables show units, freshness, and source context.
  • Confirm mobile and desktop layouts do not create text overlap or broken chart sizing when browser inspection is available.
  • For current financial, market, legal, regulatory, or source-freshness-sensitive dashboards, confirm data dates, source dates, and refresh timestamps are explicit and current.
rea-analysis-optimizationmorluto/reaOptimize REA providers and shared analysis workflows for agent context, latency, CPU, and memory while preserving correctness and Evidence. Use for analysis bottlenecks and cache or query design; CI/test-suite optimization belongs to rea-ci-tests-audit.Stars: 466kUpdated Claude CodeCodexVerifiedInstall
rea-ci-tests-auditmorluto/reaAudit REA CI and test suites for correctness, fixture drift, false passes, and measured performance; use for suite or workflow investigations and efficiency changes, rather than ordinary feature testing or a single test failure.Stars: 466kUpdated Claude CodeCodexVerifiedInstall
rea-tool-designmorluto/reaAudit, investigate, design, or change REA investigation tools, CLI/MCP contracts, provider capabilities, and Evidence semantics.Stars: 466kUpdated Claude CodeCodexVerifiedInstall
ponytail-auditDietrichGebert/ponytailQuality audit of the whole repo: bugs, security, real load, missing tests, speed, and what to delete. Most important first.Stars: 193kUpdated Claude CodeCodexVerifiedInstall