The Agent Skills Directory

[SAFE]: The skill's files consist of instructions and references for browser-based debugging tasks. No suspicious code or hidden behaviors were identified across the English and Chinese versions.
[PROMPT_INJECTION]: Analysis of the skill's attack surface identifies a potential for indirect prompt injection common to all browser-interaction tools.
Ingestion points: Untrusted data is ingested from external web pages via take_snapshot, list_console_messages, and list_network_requests (referenced in references/chrome-devtools.md).
Boundary markers: No specific boundary markers or instructions to ignore instructions within ingested data are defined in the SKILL.md files.
Capability inventory: The skill allows for significant browser-side interaction, including executing JavaScript via evaluate_script, filling forms with fill_form, and UI interaction with click and type_text.
Sanitization: The instructions do not specify any sanitization or validation of content retrieved from web pages before it is processed by the agent.

chrome-devtools