The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes local shell scripts (scripts/enumerate.sh) and Python scripts (scripts/analyze_git_security.py, scripts/generate_svg.py) as part of its analysis pipeline. These scripts use standard system tools like git, grep, and wc to analyze the local repository.
[EXTERNAL_DOWNLOADS]: The skill performs a version check by fetching a file from the author's official GitHub repository (https://raw.githubusercontent.com/pashov/skills/main/x-ray/VERSION). This is a standard update notification mechanism from a recognized vendor.
[INDIRECT_PROMPT_INJECTION]: The skill ingests external project documentation (whitepapers, specs) and source code to extract security-relevant information.
Ingestion points: Documentation files and Solidity source files are read into the agent context.
Boundary markers: The skill uses specific subagent prompts with structured extraction formats to isolate relevant facts.
Capability inventory: The skill's operations are limited to local file system analysis and writing reports to a specific x-ray/ subdirectory. It does not possess capabilities for arbitrary network exfiltration or system modification.
Sanitization: The skill relies on subagent extraction logic to filter documentation content before inclusion in the final report.

x-ray