The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill facilitates the retrieval of external web content, which introduces a surface for indirect prompt injection attacks.
Ingestion points: External web data is ingested into the agent context via the exa.searchAndContents method, which retrieves page text, summaries, and highlights (as seen in the examples in SKILL.md).
Boundary markers: The code examples do not include explicit delimiters or instructions to the agent to treat the retrieved content as data rather than instructions.
Capability inventory: The skill environment permits the use of tools such as Read, Write, Edit, Bash, and Grep across the workspace.
Sanitization: No explicit sanitization, filtering, or validation of the content retrieved from external URLs is demonstrated in the provided scripts.
[SAFE]: The skill correctly implements security best practices for credential management by using environment variables (process.env.EXA_API_KEY) to access API keys instead of hardcoding sensitive strings.
[SAFE]: References to external documentation and resources point to official Exa domains (docs.exa.ai), which are well-known and consistent with the skill's stated purpose.

exa-core-workflow-a