Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Override Markstream node renderers and add trusted custom tags across Vue, React, Svelte, and Angular using scoped or renderer-local mappings.
.claude/skills/sickn33-markstream-custom-components/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-14 | ✗→✓ | ▲ Improved | -1% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -38% | 0% |
| case-01 | ✗→✓ | ▲ Improved | -31% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -10% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -53% | 0% |
Customize specific Markstream nodes or trusted custom tags without replacing the parser or leaking global renderer state. Read references/patterns.md first.
Use to replace built-ins such as image, link, code_block, mermaid, or inline_code; render trusted tags such as thinking; or scope overrides to one renderer or app. Use parser transforms only when token or AST reshaping is required.
Before changing dependencies or source files, inspect the existing package manager and project conventions, preview the intended edits, and obtain explicit user approval.
setCustomComponents(customId, mapping); Svelte and Angular can also pass renderer-local maps.streamingComponents for parser-backed nodes and htmlComponents for sanitized attributes plus children.tsximport MarkdownRender, { type NodeComponentProps, setCustomComponents, } from 'markstream-react' import 'markstream-react/index.css' function ThinkingNode({ node }: NodeComponentProps<any>) { return <details><summary>Thinking</summary>{node.content}</details> } setCustomComponents('assistant-panel', { thinking: ThinkingNode }) export function Answer({ markdown }: { markdown: string }) { return ( <MarkdownRender content={markdown} customId="assistant-panel" customHtmlTags={['thinking']} htmlPolicy="safe" /> ) }
Treat custom HTML-like tags as trusted input only. Keep safe HTML enabled and do not pass unsanitized attributes into host components.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-14 | fail→pass | 13,128 | 10,260 | -22% | 1 | 1 | 0% | 2,357 | 2,342 | -1% | 0 | 0 | — |
case-04 | fail→pass | 13,998 | 5,079 | -64% | 1 | 1 | 0% | 2,551 | 1,586 | -38% | 0 | 0 | — |
case-01 | fail→pass | 21,031 | 10,062 | -52% | 1 | 1 | 0% | 3,689 | 2,558 | -31% | 0 | 0 | — |
case-02 | fail→fail | 13,435 | 9,633 | -28% | 1 | 1 | 0% | 2,201 | 2,418 | +10% | 0 | 0 | — |
case-03 | fail→pass | 18,118 | 11,219 | -38% | 1 | 1 | 0% | 2,938 | 2,658 | -10% | 0 | 0 | — |
case-05 | fail→pass | 11,195 | 2,417 | -78% | 1 | 1 | 0% | 1,884 | 889 | -53% | 0 | 0 | — |
case-06 | fail→pass | 11,227 | 2,849 | -75% | 1 | 1 | 0% | 1,775 | 947 | -47% | 0 | 0 | — |
case-07 | pass→pass | 8,693 | 4,053 | -53% | 1 | 1 | 0% | 1,255 | 1,112 | -11% | 0 | 0 | — |
case-08 | fail→pass | 7,026 | 4,990 | -29% | 1 | 1 | 0% | 1,201 | 1,429 | +19% | 0 | 0 | — |
case-09 | pass→pass | 12,261 | 5,941 | -52% | 1 | 1 | 0% | 1,972 | 1,498 | -24% | 0 | 0 | — |
case-10 | fail→pass | 13,493 | 5,121 | -62% | 1 | 1 | 0% | 2,220 | 1,394 | -37% | 0 | 0 | — |
case-11 | pass→pass | 6,271 | 1,603 | -74% | 1 | 1 | 0% | 1,013 | 763 | -25% | 0 | 0 | — |
case-12 | pass→pass | 11,249 | 6,742 | -40% | 1 | 1 | 0% | 1,927 | 1,650 | -14% | 0 | 0 | — |
case-13 | fail→fail | 18,059 | 9,471 | -48% | 1 | 1 | 0% | 2,875 | 2,052 | -29% | 0 | 0 | — |
case-15 | fail→fail | 9,330 | 4,766 | -49% | 1 | 1 | 0% | 1,543 | 1,474 | -4% | 0 | 0 | — |
case-16 | fail→fail | 9,598 | 1,816 | -81% | 1 | 1 | 0% | 1,545 | 864 | -44% | 0 | 0 | — |
case-17 | pass→pass | 10,884 | 4,421 | -59% | 1 | 1 | 0% | 1,708 | 1,275 | -25% | 0 | 0 | — |
case-18 | pass→pass | 6,839 | 5,309 | -22% | 1 | 1 | 0% | 1,125 | 1,476 | +31% | 0 | 0 | — |
case-19 | fail→pass | 14,411 | 9,568 | -34% | 1 | 1 | 0% | 2,726 | 2,392 | -12% | 0 | 0 | — |
case-20 | pass→pass | 9,264 | 6,235 | -33% | 1 | 1 | 0% | 1,726 | 1,634 | -5% | 0 | 0 | — |
case-21 | fail→pass | 14,521 | 6,213 | -57% | 1 | 1 | 0% | 2,642 | 1,703 | -36% | 0 | 0 | — |
case-22 | fail→pass | 11,415 | 8,648 | -24% | 1 | 1 | 0% | 1,809 | 2,003 | +11% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +50 percentage points is the difference between those two pass rates over the 22 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.