perplexity

Warn

Audited by Gen Agent Trust Hub on Mar 1, 2026

Risk Level: MEDIUMCREDENTIALS_UNSAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill documentation suggests executing a bash command to append an environment variable to the user's shell profile (~/.bashrc). This is a persistence mechanism that modifies system configuration files to store credentials.
  • [CREDENTIALS_UNSAFE]: The skill requires a PERPLEXITY_API_KEY and provides instructions to store it in a local config.json file or within shell configuration files, potentially exposing the secret if the filesystem or environment is compromised.
  • [PROMPT_INJECTION]: The skill is designed to ingest and process web search results and citations from an external API, which constitutes an indirect prompt injection attack surface.
  • Ingestion points: Web content and citations returned by the Perplexity API via the scripts/perplexity_search.sh script.
  • Boundary markers: No delimiters or instructions to ignore embedded commands in the search results are documented.
  • Capability inventory: The skill executes shell scripts and performs network operations via curl.
  • Sanitization: There is no evidence of content sanitization or validation of the data retrieved from the Perplexity API before it is processed by the agent.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Mar 1, 2026, 05:14 AM