The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes content from local markdown files in the outputs/ directory which are generated by other AI agents, creating a potential surface for instructions to be passed through data.
Ingestion points: Content is read from outputs/linkedin/, outputs/threads/, outputs/x/, and outputs/captions/ directories, as well as configuration files context/workflow-status.md and context/brand-style.md.
Boundary markers: The skill relies on specific status fields such as BLOTATO FLAG: Yes or BLOTATO FLAG: No to determine processing logic.
Capability inventory: The skill utilizes MCP tools to create posts, update schedules, and generate visual infographics on the external Blotato platform.
Sanitization: Includes a mandatory Phase 3 confirmation gate where the operator must manually approve the proposed schedule, caption, and visuals before any data is submitted to the external service.
[DATA_EXFILTRATION]: The skill transmits content gathered from local files to the blotato.com service via the provided MCP tools. This network activity is the primary functional purpose of the skill for social media publishing.

publisher