The Agent Skills Directory

[PROMPT_INJECTION]: The skill presents an indirect prompt injection surface because it references functionality for ingesting untrusted external data and has capabilities to perform write actions on the user's behalf.
Ingestion points: The troubleshooting documentation references a 'get_post' tool which allows the agent to ingest external LinkedIn content.
Boundary markers: No delimiters or explicit instructions are provided to help the agent isolate external data from its core operational instructions.
Capability inventory: The skill provides engagement tools such as 'linkedin_create_comment' and 'linkedin_create_reaction' which allow the agent to execute actions based on processed data.
Sanitization: No sanitization or validation of the external social media content is mentioned or implemented in the skill description.
[SAFE]: The skill communicates with 'mcp.publora.com' and 'api.publora.com', which are the official infrastructure endpoints belonging to the author 'publora'. These connections are necessary for the skill's intended functionality and are documented clearly.

linkedin-analytics