The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill provides instructions to clone the SpiderFoot repository from GitHub (https://github.com/smicallef/spiderfoot) and install dependencies from a requirements file. This is standard procedure for using this open-source tool.
[COMMAND_EXECUTION]: The documentation contains numerous shell commands for running the tool, including starting a web UI (python3 sf.py -l 127.0.0.1:5001) and executing CLI scans. It also provides a Python automation script using subprocess.run() to execute the tool programmatically. These are documented usage patterns for the tool's primary function.
[INDIRECT_PROMPT_INJECTION]: The skill represents a surface for indirect prompt injection as it is designed to ingest and process large amounts of untrusted data from external sources (DNS, social media, web crawlers, breach data).
Ingestion points: Data enters the context via the sf.py output (JSON/CSV) which is then processed by Python scripts described in SKILL.md and references/modules.md.
Boundary markers: None identified in the provided scripts or instructions.
Capability inventory: The skill uses subprocess.run() to execute shell commands and open()/write() for file operations.
Sanitization: The provided Python snippets perform basic JSON parsing and filtering but do not include explicit sanitization or escaping of the harvested data before potential further use.
[DATA_EXPOSURE]: The documentation mentions high-value API keys for services like Shodan, VirusTotal, and HaveIBeenPwned. It correctly advises configuring these through the Web UI or settings file, which is a standard practice for this tool and does not involve hardcoded secrets in the skill itself.

spiderfoot