The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted markdown content which serves as a vector for indirect prompt injection. Malicious instructions embedded in the input markdown could influence the agent's analysis or be rendered into the output SVG.
Ingestion points: The skill reads external markdown files as specified in SKILL.md (Step 2).
Boundary markers: No boundary markers or 'ignore' instructions are used when processing the ingested markdown.
Capability inventory: The skill possesses file-writing capabilities and can execute local commands (bun, code) via the Bash tool and scripts/render.ts.
Sanitization: There is no evidence of sanitization or escaping of the extracted markdown content before it is interpolated into the AntV infographic syntax.
[COMMAND_EXECUTION]: The skill performs shell command execution for environment setup and user convenience.
Evidence: SKILL.md executes bun run -e to check for dependencies. scripts/render.ts uses Bun.spawnSync to launch the VS Code editor for a visual preview of the generated SVG.
[EXTERNAL_DOWNLOADS]: The skill dynamically installs the @antv/infographic package if it is missing from the environment.
Evidence: SKILL.md (Step 1) triggers the install-dependency skill to download and install the @antv/infographic package from the official registry.

infographize