The Agent Skills Directory

[COMMAND_EXECUTION]: The script WebappScripts/with_server.py uses subprocess.Popen with shell=True to execute arbitrary shell commands provided by the user to start local development servers.
[DATA_EXFILTRATION]: The VulnerabilityAnalysisGemini3.md workflow instructs the agent to aggregate and send consolidated security scan findings (which may include IP addresses, endpoints, and technology stack details) to external LLM APIs for automated analysis.
[PROMPT_INJECTION]: The VulnerabilityAnalysisGemini3.md workflow establishes an attack surface for indirect prompt injection by ingesting untrusted data from multiple security scanners directly into LLM prompts for reasoning tasks.
[EXTERNAL_DOWNLOADS]: The BugBountyTool fetches bug bounty program metadata and commit history from the well-known arkadiyt/bounty-targets-data repository on GitHub.
[COMMAND_EXECUTION]: Multiple OSINT and reconnaissance workflows (e.g., Automation.md) recommend the use of automated scanning pipelines and persistence mechanisms like cron for continuous monitoring.

WebAssessment