Canvas Chart Rendering Skill (ECharts)
Enables AI to output DOJO_CHART blocks that render interactive ECharts visualizations on the Dashboard Canvas.
The skill only instructs outputting DOJO_CHART blocks, it does not perform external operations or access files/network. No permissions requested, no sensitive data handling, risk low as it relies on ECharts rendering and front-end parsing. Data source provided by user tools, least privilege details unclear. Deduction: lacks user confirmation, data flow transparency limited, external library versions and supply chain security not mentioned.
Skill provides clear protocol and templates, instructions self-consistent. However key path depends on user tools (e.g., dojo.sdk.get_stock_kline) and front-end DOJO_CHART renderer, not verified in skill file. Abnormal input (missing fields, non-numeric) not handled. No tests or execution evidence. Static review caps at 10.
Use cases clear (chart rendering), trigger conditions explicit, multiple chart templates provided. Boundaries not stated (unsupported chart types), non-fit scenarios undefined. Chinese support limited, dependency on ECharts and front-end may affect mainland China users. Environment fit evidence insufficient.
Documentation well-structured, version/author/license present, templates and examples provided. Missing install/dependency notes, changelog, maintenance responsibility and update path. License mismatch (MIT vs Apache). No FAQ or troubleshooting.
Skill likely produces chart output, but relies on front-end parsing to actually render, static review cannot verify output usability. Example has only one point; data limits defined. No comparison with alternatives. Max 7.
Only textual examples, no tests or execution evidence. Rendering done by external front-end, cannot independently reproduce. Static review caps at 5, evidence thin so lowered to 3.
- Skill depends on front-end DOJO_CHART parser, rendering will fail if not implemented.
- Data fetching relies on user tools, tool failure may produce wrong charts.
- License mismatch: SKILL.md says MIT, repository is Apache-2.0.
- No install/dependency notes; chart rendering depends on ECharts loading.
- Mainland China users should check ECharts CDN reachability.
What does this skill do, and when should you use it?
This skill teaches the AI to emit a fenced code block tagged `DOJO_CHART` containing a JSON object with data and a rendering script, which the Dashboard frontend parses and renders as an interactive ECharts chart. It includes ready-to-use templates for candlestick, line, bar, and area charts, along with a dark theme color scheme. It is part of the DojoAgents full-market AI investment assistant framework.
When the user requests a chart, the AI outputs a DOJO_CHART block in its response. The block's JSON has two fields: data (an array of points, either OHLCV objects with time/open/high/low/close/volume or simple time/value objects) and script (a JavaScript function body that calls chart.setOption(...) to render). The skill provides templates for K-line, K-line with volume, line, bar, and area charts, and mandates a dark theme palette.
- A user asks 'Draw AAPL K-line on canvas' – the agent fetches kline data via a tool and outputs a candlestick chart block.
- A user wants to compare volume across several stocks – the agent generates a bar chart.
- A user requests a trend analysis of an indicator – the agent produces a smooth line chart.
- A user wants to visualize cumulative portfolio returns – the agent uses an area chart.
What are this skill's strengths and limitations?
- Clear protocol with multiple chart templates ensure consistent, well-formed output.
- Inline dark theme colors fit common dashboard designs.
- Supports common financial chart types (candlestick, line, bar, area).
- Requires the DojoAgents Dashboard Canvas frontend to parse and render; won't work standalone.
- Data array is limited to 200 points for performance; may need downsampling for high-frequency data.
- No test suite or verification method documented; quality depends on model adherence to templates.
How do you install this skill?
This skill is part of the DojoAgents repository. To use it, install DojoAgents either via PyPI: create a virtual env with uv venv && source .venv/bin/activate then uv pip install dojoagents (no clone needed), or from source: clone the repo, create venv, and run uv pip install -e ".[dev]". The skill lives at dojoagents/skills/built_in/canvas-chart/.
How do you use this skill?
In an Agent Skills-compatible client (e.g., Claude Code), copy the skill folder into your project's skills directory or configure per client. Then prompt the agent with something like 'Draw a candlestick chart for AAPL'. The agent will fetch data and output a DOJO_CHART block. The Dashboard Canvas frontend must be available to parse and render the block.