The Agent Skills Directory

[PROMPT_INJECTION]: The skill ingests untrusted data from repository files to synthesize domain terminology, presenting a vector for indirect prompt injection.\n- Ingestion points: The agent reads content from the docs/ directory, PRD.md, README.md, and the source codebase to extract information (SKILL.md).\n- Boundary markers: The instructions do not define boundary markers or provide explicit guardrails to prevent the agent from following instructions embedded within the analyzed project files.\n- Capability inventory: The agent has the ability to read all files in the repository and write to CONTEXT.md or CONTEXT-MAP.md (SKILL.md).\n- Sanitization: No validation, escaping, or filtering of the extracted content is performed before the agent processes the information or writes it to the glossary file.

lenguaje-ubicuo