The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill installation process requires cloning a repository from an untrusted GitHub account ('sparkfinderoven') that is not recognized as a trusted vendor or well-known service.
[COMMAND_EXECUTION]: The skill provides instructions and templates for executing powerful system commands, including 'aws ec2 modify-instance-attribute' to change cloud infrastructure, 'iptables' to modify network firewall rules, and 'dd' for low-level disk imaging. While these are intended for incident response, they represent a significant security risk if misused.
[DATA_EXFILTRATION]: The suite is designed to access highly sensitive information, such as AWS IAM users, policies, and credentials, as well as system security logs and PII. It includes configurations for external integrations like Slack webhooks and SMTP servers, which could be used to exfiltrate this sensitive data.
[DYNAMIC_EXECUTION]: The skill relies on generating and executing Python and JavaScript scripts at runtime to perform tasks like dependency auditing and GDPR data flow analysis. This dynamic code generation increases the risk of accidental or malicious code execution.
[INDIRECT_PROMPT_INJECTION]: The skill possesses a large attack surface as it is designed to ingest and process untrusted external data, including source code, dependency manifests, and architecture diagrams. It lacks explicit boundary markers or sanitization logic to prevent malicious instructions embedded in these files from influencing the agent's behavior during high-privilege operations.

sparkfinderoven-claude-security-compliance-suite