The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external markdown files (brand-profile and voice-builder) to guide its content generation, which could allow malicious instructions embedded in those files to influence the agent's behavior.
Ingestion points: External files brand-profile.md and voice-builder.md are used as primary context for drafting posts and replies.
Boundary markers: Absent; there are no explicit delimiters or instructions to ignore commands within the ingested text context.
Capability inventory: The skill is restricted to drafting text content; it does not have the capability to autonomously post, execute shell commands, or perform network operations. It delegates execution to a human or separate scheduling tools.
Sanitization: Absent; the skill does not perform validation or filtering on the content of the brand or voice profile files before use.

x-growth