The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute several shell commands using npx to run the primary tool and install dependencies.
Evidence: Use of npx explorbot, npx playwright install, and npx explorbot init commands in SKILL.md and references/no-install.md.
[EXTERNAL_DOWNLOADS]: The skill fetches documentation and browser binaries from external sources.
Evidence: Downloads documentation from https://raw.githubusercontent.com/testomatio/explorbot/main/docs/ and browser binaries via Playwright's official installer.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from target websites and user-provided knowledge strings, creating a surface for indirect prompt injection.
Ingestion points: Target URLs provided to the explore command and markdown content passed to the --knowledge flag or EXPLORBOT_KNOWLEDGE environment variable.
Boundary markers: The skill documentation explicitly mentions using YAML frontmatter fences (---) to scope knowledge strings, which acts as a delimiter.
Capability inventory: The skill has the capability to execute shell commands via npx, write to the local file system (e.g., ~/.explorbot/), and perform network operations to send reports to the vendor's service.
Sanitization: It mentions that environment variables ${env.NAME} are resolved by the tool to keep secrets out of plain text knowledge files, providing a layer of variable sanitization.

explorbot-fundamentals