The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted data from external documents and links provided by the user.
Ingestion points: Identified in SKILL.md (Entrada esperada) and references/empresa.md (material bruto), where the agent reads external links or raw documents to extract institutional data.
Boundary markers: No explicit delimiters for external content are defined in the instructions.
Capability inventory: The skill is restricted to reading and writing Markdown files within the .inboundfy/context/ directory. No network exfiltration tools or shell execution capabilities were identified in the provided files.
Sanitization: The instructions explicitly mandate that the agent must confirm all extracted facts with the user before updating canonical files, significantly reducing the risk of automated obedience to embedded instructions.
[REMOTE_CODE_EXECUTION]: The skill mentions external tools like inboundfy-setup and inboundfy doctor. These appear to be vendor-provided CLI utilities for managing the local framework environment and do not constitute arbitrary remote code execution. No patterns of downloading and piping remote scripts were found.

inboundfy-contexto-institucional