The Agent Skills Directory

[COMMAND_EXECUTION]: Executes a local Python script located at scripts/data_fetcher.py to retrieve stock valuation and financial data. This script is a vendor-provided resource used for the skill's core functionality.
[COMMAND_EXECUTION]: Invokes the Google Chrome binary in headless mode or the npx utility to convert generated HTML reports into PDF files.
[EXTERNAL_DOWNLOADS]: Utilizes npx -y md-to-pdf to download and run the 'md-to-pdf' package from the NPM registry at runtime if a local Chrome installation is unavailable.
[REMOTE_CODE_EXECUTION]: The execution of npx involves fetching code from a remote repository and executing it on the local system.
[DATA_EXFILTRATION]: Uses WebFetch and WebSearch to access content from arbitrary company websites, investor relations pages, and news outlets. While intended for data collection, this involves transmitting queries to external services.
[PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection (Category 8).
Ingestion points: Phase 1 (Agent 2 and Agent 5) fetches content from third-party news sites and company websites via WebFetch.
Boundary markers: The prompts for the subagents do not include instructions to treat the fetched content as untrusted or to ignore embedded instructions.
Capability inventory: The agent has the capability to write files to the local filesystem (HTML/PDF) and execute shell commands in Phase 4.
Sanitization: No explicit sanitization or filtering of the retrieved web content is performed before it is summarized by the model or included in the final HTML report.

analyze-stock