The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill provides references and installation links for CLI tools hosted on the official GitHub repository for Inference.sh.
[COMMAND_EXECUTION]: The skill employs the belt command-line utility to interact with remote research applications. The execution environment is restricted via the allowed-tools configuration to only permit commands starting with the belt prefix.
[DYNAMIC_EXECUTION]: The skill uses a remote execution environment (infsh/python-executor) to run Python scripts defined within the skill body. This is used to generate competitive positioning maps and other visualizations.
[INDIRECT_PROMPT_INJECTION]: The skill ingests and summarizes content from external websites, search engine results, and third-party reviews, which introduces a potential surface for instructions embedded in those external sources to influence the agent's output.
Ingestion points: Processes external content through infsh/agent-browser, tavily/extract, and tavily/search-assistant as seen in SKILL.md.
Boundary markers: No explicit delimiters or instructions are used to distinguish untrusted external data from the skill's primary instructions.
Capability inventory: The skill possesses capabilities for network research, browser automation, and file creation (e.g., saving visualization images).
Sanitization: There is no evidence of sanitization or filtering for the external content before it is processed by the model.

competitor-teardown