The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill provides numerous links to external resources, including educational videos on YouTube, source code repositories on GitHub, and documentation from well-known technology organizations and academic institutions. References include official content from Microsoft, Google, Anthropic, OpenAI, Stanford University, MIT, and Carnegie Mellon University. These links are for informational and educational purposes.\n- [COMMAND_EXECUTION]: The skill includes a shell script scripts/run.sh that facilitates the execution of the agent via a Python wrapper. The script handles directory path resolution and argument passing to ensure the skill runs within its intended environment. This is a standard operational pattern for this type of agent skill.\n- [INDIRECT_PROMPT_INJECTION]: The skill includes instructions for the agent to verify the availability and currency of external URLs. This involves the agent accessing and potentially processing content from external websites, which constitutes an indirect prompt injection surface. An attacker could theoretically place malicious instructions on a website that the agent visits during its verification process.\n
Ingestion points: User tasks (queries) and the content of external URLs analyzed during the link verification step defined in the protocol.\n
Boundary markers: No specific delimiters or "ignore embedded instructions" warnings are defined for processing external data.\n
Capability inventory: The skill utilizes subprocess execution to launch the agent via scripts/run.sh, and the agent likely employs network tools (e.g., curl, browser) to verify curated links.\n
Sanitization: The skill does not define specific sanitization or filtering logic for the content it retrieves from external sources.

specialized-ai-learning-curator