The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The environment setup scripts fetch essential components from external sources.
Downloads the uv package manager installer from the well-known astral.sh domain.
Clones the ida-domain source repository from the official HexRaysSA organization on GitHub.
Installs standard libraries including playwright, puppeteer, pandas, and openpyxl from public registries.
[REMOTE_CODE_EXECUTION]: The setup process involves executing an external installation script.
The installer for the uv tool is piped directly to the shell via curl -LsSf https://astral.sh/uv/install.sh | sh. While this is an established installation method for a well-known service, it represents a high-privilege execution pattern.
[DYNAMIC_EXECUTION]: The skill provides an IDA Pro script execution environment (run.py).
It dynamically assembles Python code by indention and wrapping provided snippets in boilerplate, then executes the result via a subprocess. This dynamic generation and execution of code is the core intended functionality for the IDA scripting skill.
[INDIRECT_PROMPT_INJECTION]: The scraping modules ingest and process content from various external websites, creating a surface for indirect instructions.
Ingestion points: Scraping scripts for Trustpilot (tpscraper.js), Google Ads (scrape-ads-playwright.cjs), and general web crawling (batch_crawler.py) ingest arbitrary text from the internet.
Boundary markers: The scripts do not implement specific delimiters or 'ignore' instructions to isolate scraped data from the agent's logic.
Capability inventory: The skill has permissions to execute shell commands, perform network operations, and write to the local file system, which defines a high capability tier for potentially untrusted data to influence.

scraping-specialist