The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the firecrawl CLI tool to manage monitoring tasks, such as creating monitors (create), listing results (list), and viewing specific checks (check). It also supports writing check outputs to the local file system via the -o or --output flags.
[EXTERNAL_DOWNLOADS]: The skill invokes npx firecrawl, which downloads the Firecrawl package from the npm registry. As Firecrawl is the skill author, this is a standard vendor-provided dependency required for the skill's functionality.
[DATA_EXFILTRATION]: The skill can send notifications to external webhooks and email addresses via the --webhook-url and --email flags. This functionality is configured using explicit, user-provided parameters as part of the monitor's intended notification feature.
[PROMPT_INJECTION]: The skill has a surface for indirect prompt injection because it ingests untrusted data from external websites for the purpose of change detection and AI-based judging.
Ingestion points: External content is pulled from target URLs (--page, --scrape-urls, --crawl-url) or from broad web search results in 'web monitors' (--queries).
Boundary markers: The skill documentation does not specify the use of delimiters or 'ignore' instructions to isolate processed web content within the agent's context.
Capability inventory: The agent can execute shell commands (firecrawl) and perform local file system writes.
Sanitization: The tool utilizes an 'AI judge' to filter out formatting and timestamp noise, but no security-focused sanitization or filtering of the web content is specified to prevent malicious instructions from influencing the agent.

firecrawl-monitor