The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the user to install the @tiny-fish/cli package from the NPM registry to enable its core functionality.
[COMMAND_EXECUTION]: The skill uses shell commands to check environment status (which tinyfish, tinyfish --version), manage authentication (tinyfish auth login), create directories (mkdir), and execute web-browsing tasks (tinyfish agent run). It also employs shell backgrounding and the wait command to manage parallel execution of discovery and reading tasks.
[INDIRECT_PROMPT_INJECTION]: The skill's primary function is to ingest and process untrusted data from the live web, creating a potential vector for malicious content to influence agent behavior.
Ingestion points: External URLs are accessed via tinyfish agent run during both the discovery pass and the reading pass in SKILL.md.
Boundary markers: The skill uses specific prompts requesting JSON output to structure the extraction, but it does not implement explicit boundary markers or delimiters to isolate untrusted web content from the agent's instructions.
Capability inventory: The skill has the capability to execute shell commands (via the tinyfish CLI), create directories, and write multiple markdown and JSON files to the local file system.
Sanitization: There is no evidence of explicit content sanitization, filtering, or validation of the data extracted from web pages before it is processed by the agent to build the knowledge base.

kb-builder