The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes multiple external JSON inputs, including report_payload, chart_spec, and supporting_metrics. These are processed to generate HTML strings. This workflow creates an attack surface where instructions embedded within the source data could influence the agent's behavior or lead to the generation of malicious HTML content.
Ingestion points: Data enters the agent context via the report_payload, chart_spec, and supporting_metrics inputs mentioned in the SKILL.md.
Boundary markers: The instructions do not define specific delimiters or warnings to ignore instructions embedded within the processed payloads.
Capability inventory: The skill is primarily instructional and focused on HTML assembly; no subprocess calls or file-writing scripts were provided for analysis.
Sanitization: There are no explicit instructions for sanitizing or escaping the incoming data before it is interpolated into the HTML output.
[NO_CODE]: The skill consists entirely of markdown instructions and metadata. No scripts, binaries, or configuration files were included for analysis. The security posture is based on the instructions provided to the agent.

visual-explainer