The Agent Skills Directory

[COMMAND_EXECUTION]: The skill invokes local Python scripts located within the plugin's internal directory (${CLAUDE_PLUGIN_ROOT}/scripts/connectors/) to perform data analysis tasks. Specifically, it uses gdelt.py for news mentions, ledger.py for tracking historical data changes, and tavily.py for checking AI citations.
[EXTERNAL_DOWNLOADS]: The skill references Cloudflare Radar (radar.cloudflare.com) as a source for industry-wide AI crawler benchmarks. This is a well-known and reputable service used for informational purposes.
[INDIRECT_PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection due to its core functionality of processing external data.
Ingestion points: The skill ingests data from backlink CSV exports, web analytics (GA4/GSC) reports, and raw server logs provided by the user or fetched via tools.
Boundary markers: The instructions explicitly warn the agent to treat all external or pasted content as untrusted input and strictly forbid executing any instructions found within that data.
Capability inventory: The skill has access to the WebFetch tool and the ability to execute internal Python scripts to process data.
Sanitization: No explicit technical sanitization (like JSON schema validation) is mentioned, relying instead on high-level agent behavioral instructions to prevent accidental instruction following.

offsite-signal-analyzer