The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill instructions in SKILL.md use npx -y mcporter to download and execute code from the NPM registry at runtime. Executing unpinned external packages via npx is a security risk as it allows for the execution of arbitrary code if the package is compromised.
[CREDENTIALS_UNSAFE]: The documentation in references/exa-tools.md explicitly recommends passing the exaApiKey as a query parameter in the URL (e.g., ?exaApiKey=YOUR_EXA_API_KEY). Sensitive credentials in URLs are often exposed in web server logs, proxy logs, and browser history.
[COMMAND_EXECUTION]: The skill provides examples of command-line execution using npx and shell variables to interact with the Exa API. If user-controlled input is incorporated into these commands without strict sanitization, it could lead to command injection.
[EXTERNAL_DOWNLOADS]: The skill makes network requests to https://mcp.exa.ai/mcp to retrieve tool configurations and search results. While Exa is a known service, the skill depends on the security of this external infrastructure.
[PROMPT_INJECTION]: The skill ingests untrusted data from the web via search tools, creating a surface for indirect prompt injection.
Ingestion points: Search results and code snippets retrieved via web_search_exa and get_code_context_exa enter the agent context (File: SKILL.md).
Boundary markers: No specific delimiters or instructions to ignore instructions within the retrieved content are present in the parameter templates.
Capability inventory: The skill has the capability to execute shell commands via npx and perform further network operations.
Sanitization: There is no evidence of sanitization or filtering of the content returned from the external search API before processing.

exa-search