The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted repository data to determine installation plans, creating a surface for indirect prompt injection.
Ingestion points: Repository files accessed via rg and cat commands, including package manifests like package.json, Cargo.toml, and source code.
Boundary markers: No specific boundary markers or instructions to ignore embedded commands within the analyzed data are defined in the workflow.
Capability inventory: The skill reads project files and can write or update configuration files and routers within the .claude/skills/ directory.
Sanitization: No specific content sanitization or validation of the ingested repository data is described before the classification process.

agent-sort