The Agent Skills Directory

[SAFE]: A thorough review of all skill components, including the Python scripts (intent_analyzer.py, outline_generator.py, platform_insights.py, trend_analyzer.py), confirms that the skill performs only local data processing. There are no external network calls, file system modifications, or system command executions present in the code.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied inputs such as topic and target_audience to generate research reports. While this establishes a theoretical ingestion surface for indirect prompt injection, the risk is negligible due to the absence of exploitable capabilities within the skill's environment.
Ingestion points: Untrusted data enters the agent context through the topic, target_audience, and additional_requirements fields defined in the skill input format.
Boundary markers: Explicit boundary markers or instructions to ignore embedded commands are not present in the provided scripts.
Capability inventory: All scripts utilize standard libraries (re, json, datetime, typing) for string manipulation and logic. No subprocess, network, or file writing tools are present.
Sanitization: User input is analyzed using regular expressions for classification purposes, but no security-focused sanitization is implemented, which is acceptable given the low-impact nature of the output.

content-trend-researcher