html
Create inline HTML documents with CSS and embedded JavaScript in Pioneer messages, including interactive calculators, charts, and small working interfaces. Do not use for ordinary application edits merely involving HTML.
- License 许可
- MIT
- 最近更新
- 2026-11-12
- Stars
- ★ 21
- Forks
- 3
01 前言 Front Matter
- name
- html
- slug
- html
- owner
- pioneer
- version
- 0.1.0
- description
- Create inline HTML documents with CSS and embedded JavaScript in Pioneer messages, including interactive calculators, charts, and small working interfaces. Do not use for ordinary application edits merely involving HTML.
- catalog-hide
- true
- user-invocable
- true
- implicit-invocation
- required
- disable-model-invocation
- false
02 资源文件 7
- 许可证LICENSE1.0 KB
- 参考文档references/charts-and-data.md2.0 KB
- 参考文档references/diagrams.md4.4 KB
- 参考文档references/documents-and-plans.md2.0 KB
- 参考文档references/interactive-content.md1.9 KB
- 参考文档references/theme-and-layout.md1.6 KB
- 参考文档references/tools-and-runtime.md2.1 KB
03 SKILL.md 正文
Pioneer Inline HTML and JavaScript
Inline HTML supports CSS and embedded JavaScript that runs inside the document. Use JavaScript for working calculators, filters, tabs, scenario controls, and charts that respond to user input. Keep reading-focused documents static when interaction adds no value.
Use inline HTML to make relationships, comparisons, data, or interactive calculations easier to understand. Choose the form from the user's question: a document for reading, aligned rows for comparison, a diagram for relationships, a chart or table for data, or a small working interface for exploration. Plain text is sufficient for simple answers.
Workflow
- Prepare one complete HTML document with ready-to-run CSS and JavaScript. No TypeScript, TSX, or required build step.
- When
html_previewis available and image input is supported, call it using the callable name in the current runtime catalog. - For a successful preview, inspect all six real images and each frame's diagnostics: light at 360, 768, 1024 CSS px, then dark at the same widths. Correct clipping, overflow, contrast, layout, script errors, and truncated captures; preview again as needed.
- Publish the finished version with permitted
html_render. It already saves and publishes the document; do not addartifact_register. Iterate through preview rather than publishing drafts. - Add only necessary text that complements the HTML. Follow the current answer delivery and review rules; HTML does not bypass them.
If preview is unavailable, or image capability is explicitly unsupported or unconfirmed, do not repeat a known unsupported preview without changed conditions. Do not claim visual verification. You may publish with permitted html_render while clearly telling the user that visual verification was unavailable. If visual verification is necessary for a reliable result, provide a text answer instead. The document must still support both themes.
Preview checks initial rendering in both standard Pioneer themes, not every interaction or every future custom palette. Exercise interactions when an authorized tool supports it; never claim actions were tested from static screenshots alone.
Host contract
Adapt to message width and let content determine height. Do not add an outer decorative page shell, horizontal page padding, or viewport-height layouts that interfere with height measurement. Contain intentionally wide tables or diagrams locally.
Use host-provided --pioneer-* variables for primary styles. Support both light and dark and live theme updates. Never redefine reserved host tokens with a fixed palette. The application supplies the actual theme; prefers-color-scheme alone cannot determine the user's Pioneer choice. JavaScript-rendered Canvas/charts must reread computed tokens and redraw on pioneer-theme.
Internal controls must work within the document. HTML buttons cannot call agent tools, send chat messages, or grant permissions. Do not promise persistent DOM state: the webview may be recreated.
Read the matching references
- Always read references/theme-and-layout.md before styling.
- Read references/tools-and-runtime.md for tool inputs, diagnostics, publication, and environment limits.
- Read references/documents-and-plans.md for documents, comparisons, plans, and roadmaps.
- Read references/diagrams.md for architecture, sequence, topology, and state.
- Read references/charts-and-data.md for charts, tables, and quantitative claims.
- Read references/interactive-content.md for calculators and working interfaces.
In native runtime, use the exact skill:<id> from Internal Skill References; resolve package paths under the returned skill_asset_root. In Codex, Claude, and OpenCode, system:pioneer/html means the exported pioneer-system-html skill; resolve references relative to its loaded directory. Use only tools actually present in the runtime catalog. Skills do not grant permissions.
04 同仓库更多 Skills
threads
Use this skill to find and read accessible conversations, follow saved sources and child threads, read saved tool results, create a conversation, send a message, or start a permitted thread turn.
tasks
Use this skill when Pioneer should create, inspect, update, reschedule, pause, resume, finish, cancel, troubleshoot, or configure delivery for durable tasks, scheduled tasks, recurring tasks, interval tasks, cron jobs, background work, or existing task state.
subagents
Use this skill when Pioneer should delegate current-turn work to attached task-backed subagents, coordinate parallel child agents, wait for child results, review/revise/accept/cancel/detach attached work, or synthesize accepted child results into the parent answer.
runtime-diagnostics
Investigate Pioneer Gateway or desktop CPU, memory, slow work, queues, and runtime stalls using process metrics, local snapshots, bounded native profiles, and exact release symbols. Use for Pioneer performance diagnosis and interpreting its diagnostic artifacts.
memory
Use this skill when Pioneer should proactively use durable memory or recalled context: decide whether memory can improve a turn, answer from remembered user/project facts, request memory tools, search/list/get stored memories, save durable preferences or project decisions, forget memories, audit or clean up memory, or recover from memory tool failures.
browser
Use agent-browser to interact with websites, inspect rendered pages, fill forms, extract data, capture screenshots or PDFs, and test web interfaces. Includes browser-backed QA and CDP-enabled Electron app workflows.