The Agent Skills Directory

[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by ingesting and processing data from external AI search engines (Doubao, Kimi, DeepSeek).
Ingestion points: SKILL.md (Step 4.2) directs the agent to read and perform deep analysis on AI search results stored in output/search_results.json.
Boundary markers: The skill provides a mitigation step in SKILL.md (Step 1.3), requiring an AskUserQuestion interaction for the user to confirm or modify generated questions before the search process begins.
Capability inventory: The skill possesses the capability to execute shell commands for its local scripts, write to the file system, and perform network operations via requests.
Sanitization: The scripts/lib/report_template.py file implements data sanitization using html.escape (aliased as _esc) to prevent potential XSS when rendering external content in the interactive HTML report.
[EXTERNAL_DOWNLOADS]: The skill requires the installation of the requests library via pip3 install requests, which is a standard and well-known dependency for Python-based network operations.
[COMMAND_EXECUTION]: The workflow relies on executing internal Python scripts (geo_search.py, geo_analyze.py, and geo_report.py) to manage API interactions, perform data aggregation, and generate the final report.

geo-analyzer