The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill requires the agentcash package from the npm registry, which is managed and executed via the npx package runner.
[COMMAND_EXECUTION]: The skill utilizes the npx agentcash@latest fetch command to communicate with neural search and web scraping endpoints.
[DATA_EXFILTRATION]: Queries and URLs are sent to https://stableenrich.dev for the purpose of search and content extraction. This activity is expected based on the skill's stated purpose and targets the service associated with the author.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it reads and processes data from external, potentially untrusted websites.
Ingestion points: Scraped content and search results returned from the Exa and Firecrawl APIs.
Boundary markers: The provided instructions do not include specific boundary markers or instructions to ignore embedded commands in the scraped data.
Capability inventory: The agent's capabilities in this context are primarily limited to summarizing and synthesizing information from the search results.
Sanitization: There is no mention of filtering or sanitization steps performed on the ingested web content before it is presented to the agent.

web-research