The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides bash script snippets intended for the agent to execute, including curl operations to check script sizes and verify page content on the 'treg.to' domain.
[DATA_EXFILTRATION]: The instructions involve network operations targeting 'treg.to', which is not a whitelisted or well-known domain. These operations are used for probing and verification but technically involve sending requests to an external endpoint.
[INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to ingest and analyze data from external sources and database queries, which could potentially contain malicious instructions if the source is compromised.
Ingestion points: The agent is directed to process the output of curl requests (JS files and HTML page content) and the results of SQL queries targeting the org and adconversion tables in SKILL.md.
Boundary markers: No specific delimiters or warnings to ignore embedded instructions are provided in the skill for these ingestion points.
Capability inventory: The skill utilizes shell execution capabilities (curl, grep, wc) and database access (select queries).
Sanitization: There are no explicit instructions for sanitizing or escaping the data retrieved from the network or database before processing.

ads-conversion-tracking