The Agent Skills Directory

[COMMAND_EXECUTION]: The skill invokes local Python scripts (scripts/dfs_client.py and scripts/report_publish.py) to perform data lookups and generate client deliverables. These scripts are executed with arguments derived from user input (brand names, topics) and internal report paths.
[INDIRECT_PROMPT_INJECTION]: The skill ingests and analyzes content from external, untrusted sources which could contain malicious instructions designed to influence the agent's behavior.
Ingestion points: The webfetch tool is used to retrieve robots.txt and page content from target domains. Additionally, scripts/dfs_client.py retrieves brand mentions and topic content from external web indices.
Boundary markers: The instructions do not specify the use of delimiters or 'ignore' instructions when processing the content fetched from the web.
Capability inventory: The agent has the capability to execute shell commands (via the Python scripts) and write files to the filesystem (e.g., PERPLEXITY-<domain>-<date>.md).
Sanitization: There is no evidence of sanitization or filtering applied to the external content before it is processed by the LLM.

perplexity-optimization