The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from multiple external project files and user-supplied inputs, creating a surface for indirect prompt injection.
Ingestion points: The skill explicitly reads from several files including .inboundfy/context/empresa.md, .inboundfy/estrategia.md, .inboundfy/context/marca-voz.md, .inboundfy/context/publico.md, .inboundfy/context/proibicoes.md, .inboundfy/context/glossario.md, and .inboundfy/pipeline.md as specified in SKILL.md.
Boundary markers: No explicit delimiters, XML tags, or "ignore embedded instructions" warnings are utilized to isolate the content of these files from the agent's execution logic.
Capability inventory: The skill's capabilities are limited to text generation, file updates (Brief/ID preservation), and calling other internal skills (inboundfy-setup, inboundfy-anti-slop). No dangerous capabilities like network exfiltration, arbitrary shell execution, or privilege escalation were found.
Sanitization: While the skill invokes inboundfy-anti-slop and inboundfy-copy-edicao for editorial review, there is no technical sanitization or schema validation mentioned to prevent processing malicious instructions embedded in the project files.

inboundfy-growth-anuncios