The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes content from documentation files (AGENTS.md, CLAUDE.md) and any linked or imported documents to perform refactoring and contradiction detection. This ingestion of external, untrusted data creates a surface for indirect prompt injection.
Ingestion points: The skill reads AGENTS.md or CLAUDE.md from the repository root, along with any documents linked or imported within those files (SKILL.md).
Boundary markers: No explicit boundary markers or instructions to ignore embedded prompts are provided in the skill instructions to separate the data being processed from the agent's logic.
Capability inventory: The skill is capable of reading files, orchestrating subagents, and proposing/creating new file structures (SKILL.md).
Sanitization: The instructions do not define any sanitization or validation steps for the content found in the processed documentation files.

agents-review