The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill facilitates the creation and attachment of tool definitions (schemas and descriptions) to agent configurations. This introduces an attack surface where malicious tool definitions could influence agent behavior.
Ingestion points: The create-ai-tool MCP tool accepts a schema and description which may originate from untrusted sources or user input.
Boundary markers: The instructions do not specify the use of delimiters or warnings to ignore instructions embedded within these configuration fields.
Capability inventory: The skill utilizes high-privilege MCP tools including create-ai-tool, update-ai-config-variation, and get-ai-config to modify agent capabilities.
Sanitization: There is no mention of validation or sanitization of the JSON schemas or descriptions provided during the creation process.
[SAFE]: The skill is authored by 'launchdarkly' and references official documentation and resources associated with the LaunchDarkly platform.
[SAFE]: No hardcoded credentials, malicious remote code execution patterns, or persistence mechanisms were detected in the instructions or provided code snippets.

tools