The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill methodology involves ingesting and processing data from external, untrusted sources such as OSINT APIs and web content during the crawling phase.
Ingestion points: Data is fetched from OSINT engines like FOFA (SKILL.md, Phase 2) and through website crawling (SKILL.md, Phase 3).
Boundary markers: The instructions do not define specific delimiters or instructions to treat the ingested external data as untrusted text within the agent's execution context.
Capability inventory: The skill utilizes shell-based reconnaissance tools (subfinder, amass, ksubdomain, curl, dig) and requires network access (references/enumeration-tools.md).
Sanitization: The documentation does not specify sanitization or validation routines for data retrieved from external sources before it is processed by the agent or included in command pipelines.
[COMMAND_EXECUTION]: The reference documentation in references/enumeration-tools.md contains shell script examples that interpolate variables derived from external discovery into command strings (e.g., using curl and dig on discovered subdomains). While standard for security workflows, this pattern represents a potential injection point if the source list of subdomains is maliciously crafted.

subdomain-deep