The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes data retrieved from a local knowledge base and external web search results, which creates a potential surface for indirect prompt injection if those sources contain malicious instructions.
Ingestion points: The skill retrieves data using gbrain get and gbrain query in SKILL.md and processes synthesized web content from Perplexity.
Boundary markers: The skill instructions suggest a structured prompt template (Topic: <topic> / Brain context: <context>), which helps delimit data from instructions.
Capability inventory: The skill has the ability to write to local storage (gbrain put) and make network requests via curl to the Perplexity API.
Sanitization: There is no explicit mention of sanitizing or escaping the content interpolated into the API prompts.
[EXTERNAL_DOWNLOADS]: The skill performs network operations to reach the Perplexity API at api.perplexity.ai. This is a well-known service used for the skill's primary research purpose.
[CREDENTIALS_UNSAFE]: The skill requires a PERPLEXITY_API_KEY. The documentation correctly instructs the user to provide this via an environment variable or a .env file, which is a standard and safe method for secret management.

perplexity-research