The Agent Skills Directory

[PROMPT_INJECTION]: The skill presents an indirect prompt injection surface because it processes untrusted data from the web.
Ingestion points: Data enters the context via the get_observations function, which retrieves extension captures including console logs, network requests, and DOM content.
Boundary markers: The documentation lacks specifications for delimiters or 'ignore embedded instructions' warnings for the log data.
Capability inventory: The skill documentation references a finding skill that converts these observations into insights, providing a path for injected content to influence agent logic.
Sanitization: There is no mention of sanitization, filtering, or validation of the captured raw web data.
[NO_CODE]: The skill is composed entirely of markdown documentation and metadata and does not provide or execute any code files, scripts, or package installations.

observe