The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute various shell commands, including find, grep, git, and cp, to locate design tokens, style guides, and manage file backups. It also executes local scripts provided within the skill package: scripts/survival-check.py and scripts/render.sh.
[EXTERNAL_DOWNLOADS]: The render.sh script utilizes a headless Chrome or Chromium browser to fetch and render content from user-provided URLs to verify interface changes through screenshots.
[DATA_EXFILTRATION]: The skill performs network requests to external domains (via headless Chrome in render.sh) based on user-provided input. While intended for UI verification, this represents a network operation to non-whitelisted domains.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user-supplied files, directories, and URLs. Because it has the capability to modify the file system and execute shell commands, it presents a surface for indirect prompt injection if the source data contains malicious instructions.
Ingestion points: Target files, directory structures, and user-provided URLs specified in SKILL.md.
Boundary markers: None; the skill does not implement specific delimiters or instructions to ignore embedded commands in the processed data.
Capability inventory: File system read/write, shell command execution (find, grep, git, cp), and local script execution (scripts/survival-check.py, scripts/render.sh).
Sanitization: There is no mention of sanitization or filtering of the processed content to prevent instruction injection.

ai-sanitize