The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests and processes untrusted content from the open web, which could be crafted to include malicious instructions for the agent.
Ingestion points: The skill fetches external web pages (home, about, and target URLs) using a web retrieval tool and collects brand mentions and search results via scripts/dfs_client.py.
Boundary markers: There are no specific delimiters or instructional guardrails defined to prevent the agent from following instructions potentially embedded in the fetched web content.
Capability inventory: The skill utilizes shell access to execute Python scripts (dfs_client.py, report_publish.py) and has file system access to write multi-page reports.
Sanitization: The instructions do not define any sanitization or validation steps for the content retrieved from external sources before it is processed by the agent.
[COMMAND_EXECUTION]: The skill uses the shell to execute local Python utility scripts as part of its workflow.
It invokes python scripts/dfs_client.py to pull search engine data and python scripts/report_publish.py to generate final deliverables, using arguments derived from user input.

gemini-optimization