The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to fetch and analyze social media mentions and brand-related content from the Octolens API.
Ingestion points: External data is ingested from the https://app.octolens.com/api/v1/mentions endpoint into the agent's context through scripts like fetch-mentions.js and query-mentions.js (found in the scripts/ directory).
Boundary markers: The skill instructions lack explicit guidance on using boundary markers or delimiters when processing the body field of retrieved mentions, which may contain attacker-controlled text from platforms like Twitter or Reddit.
Capability inventory: The skill utilizes the Node Read tool and bundled Node.js scripts to perform network requests and process data.
Sanitization: There is no evidence of content sanitization or instruction-filtering for the ingested social media text before it is presented to the agent for analysis.
[EXTERNAL_DOWNLOADS]: The scripts/README.md documentation provides setup instructions for Linux environments that involve downloading and executing a configuration script from Nodesource (deb.nodesource.com), an established service for Node.js distributions.

octolens