data-analysis-for-feishu

Name/description (Feishu-oriented ECharts visualizations) align with included files: data parsing, chart JSON generation, headless rendering (pyppeteer) and auto-analysis. Required capabilities (pandas, openpyxl, pyppeteer, optional Feishu JSON inputs) are consistent with the described feature set (Excel/CSV parsing, screenshot rendering, Feishu card JSON).

SKILL.md and scripts instruct only on parsing local data or user-provided Feishu JSON, generating chart options, and rendering screenshots or card JSON. The runtime does fetch ECharts from a CDN for rendering and the headless browser downloads Chromium on first run — both are justified by the rendering requirements but are external network operations. The test script uses subprocess.run to exercise local commands; its commands are static in the repository. No instructions ask the agent to read unrelated system files, environment secrets, or to transmit data to unknown endpoints.

There is no formal install spec; the repo contains a requirements.txt listing pandas, openpyxl, requests, and pyppeteer. This is proportionate to the task, but pyppeteer will download a Chromium binary on first use (large download, network access). ECharts is loaded from jsdelivr CDN during rendering (external dependency). These network downloads are expected for local rendering but raise the usual risks of third-party resources (supply-chain/network availability).

The skill does not request environment variables, API tokens, or config paths. Feishu integration is supported but implemented as accepting Feishu JSON inputs (bitable/sheet data) rather than requiring stored Feishu credentials — so there is no unexpected credential request in the package.

Skill flags are standard (not always: true). The package does not attempt to modify other skills or system-wide configs. It runs as a normal skill with no elevated persistent privileges.