The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the Bash tool to execute local development commands, including the studio CLI for WordPress management, node for running helper scripts (scripts/run.mjs), and ImageMagick tools (identify, convert) for cropping and comparing screenshots.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and processes styledHtml captured from external source websites to inform its design decisions.
Ingestion points: The styledHtml field within outputDir/sections/<slug>.json, as well as configuration data in <outputDir>/asset-triage.json.
Boundary markers: The instructions lack specific delimiters or instructions for the agent to ignore potentially malicious content embedded within the source HTML.
Capability inventory: The skill has access to powerful tools like Bash (command execution), Write and Edit (filesystem access), and the studio CLI (WordPress database/site modification).
Sanitization: There is no evidence of sanitization or filtering applied to the captured source content before it is used by the AI model to generate WordPress block markup.

match-section