The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill's HTML templates load the ECharts visualization library from jsDelivr, a well-known and trusted content delivery network. This is a standard practice for including browser-based charting capabilities.
[COMMAND_EXECUTION]: The skill utilizes the opencli utility to fetch public data from social media platforms. This activity is restricted to the specific tool mentioned in the skill's configuration and is necessary for its primary function of public opinion monitoring.
[PROMPT_INJECTION]: The skill processes untrusted content from social media comments to perform sentiment analysis.
Ingestion points: Comments are ingested from social media platforms via opencli commands as described in references/platform-guide.md.
Boundary markers: The analysis prompt in references/analysis-guide.md uses clear section headers to separate instructions from the comments provided in JSON format.
Capability inventory: The skill has the ability to run shell commands (opencli) and write HTML files to the local system.
Sanitization: The workflow in SKILL.md includes steps for data cleaning (Step 3.1) and author de-identification (Step 4.4) before inclusion in the final report, mitigating risks associated with untrusted data processing.

sentiment-monitor