The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to read, summarize, and act upon output files generated by scheduled tasks and external SEO tools (e.g., geogrid scans, citation audits). This creates a surface where malicious or malformed content in these external reports could influence the agent's behavior.
Ingestion points: Files located in briefs/{brand}/{location}/scans/ and briefs/{brand}/{location}/reports/.
Boundary markers: The instructions do not specify the use of delimiters or 'ignore embedded instructions' warnings when processing these files.
Capability inventory: The skill has the ability to write to the filesystem, execute the LocalSEOData tool suite, and potentially interact with Slack.
Sanitization: There are no mentioned processes for sanitizing or validating the content of external reports before they are processed by the agent.
[COMMAND_EXECUTION]: During the 'First Run Setup,' the skill is instructed to automatically execute a comprehensive list of tools (including local_audit, geogrid_scan, and citation_audit) without requesting specific user permission for the tools to run. This reduces user oversight over the agent's initial tool interactions.

brief