---
name: render-insight
description: Render an SVG card from explicit title, claim and evidence-boundary text. Use for reusable research visuals, not to invent or verify facts.
compatibility: Requires Python 3.10 or later; no network or third-party libraries are needed by the renderer.
metadata:
  lri.profile: "lri.capability/v0.1"
  lri.manifest: "assets/capability.json"
---

# Render an insight card

## Inputs
JSON containing `title`, `claim` and `boundary`, all non-empty strings.

## Procedure
1. Confirm the claim's evidence boundary; read `references/BOUNDARIES.md` when uncertain.
2. Run `python3 scripts/render.py --input INPUT.json --output OUTPUT.svg`.
3. Inspect the three fields in the SVG and report the exact output path.

## Acceptance and recovery
Generate a valid self-contained SVG. Missing inputs and existing output paths are errors. Select a new destination instead of overwriting an unrelated file. Rendering does not establish factual truth.

## Optional profile
`assets/capability.json` is an LRI proposal, not a public Agent Skills standard. A host without it can follow the procedure above. Metadata grants no permission. This example implements only local SVG output, not MP4, PowerPoint, iframe hosting or remote hardware.