The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the infsh command-line interface to perform market research tasks. Access is restricted via the allowed-tools frontmatter to the infsh command only.
[EXTERNAL_DOWNLOADS]: References installation instructions for the infsh CLI located at raw.githubusercontent.com/inference-sh/skills. It also suggests adding related modules using npx skills add from the same organization.
[REMOTE_CODE_EXECUTION]: Includes a Python script executed via the infsh/python-executor tool. This script uses matplotlib to generate a positioning map image. The code is a static template provided within the skill instructions and does not incorporate unsanitized external inputs into the code structure.
[SAFE]: The skill processes untrusted data from external websites (via agent-browser and search tools). While this represents a surface for indirect prompt injection, it is the primary intended function of a research skill.
Ingestion points: Competitor websites, G2/Capterra reviews, and search engine results.
Boundary markers: None explicitly defined in the provided examples.
Capability inventory: Web browsing (agent-browser), file writing (python-executor), and network search (tavily, exa).
Sanitization: Standard tool-level sanitization is assumed; the skill itself does not implement additional filtering for retrieved content.

competitor-teardown