The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill downloads and executes the web-forager and ddgs (DuckDuckGo Search) packages from public registries using the uvx and uv run commands.
[EXTERNAL_DOWNLOADS]: The skill fetches web content through the well-known Jina Reader service via https://r.jina.ai/ to convert web pages into markdown format for processing.
[COMMAND_EXECUTION]: The skill utilizes shell commands (uvx, uv run, and curl) to execute research tools and perform network operations.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and processes untrusted data from external websites encountered during research.
Ingestion points: Raw web page content fetched via the web-forager fetch tool or the Jina Reader API (SKILL.md).
Boundary markers: The instructions do not define specific delimiters or instructions to the agent to ignore potentially malicious embedded instructions within the fetched content.
Capability inventory: The agent has capabilities for network access (curl, uvx) and command execution (uvx, uv run) which could be targeted by sophisticated injections.
Sanitization: There is no explicit requirement for the skill to sanitize or filter the content of fetched pages before synthesizing them into the final report.

deep-research