The Agent Skills Directory

[COMMAND_EXECUTION]: The skill documents various injection payloads, including SQL injection strings containing commands like pg_sleep and NoSQL injection strings using $regex operators, intended for testing target systems.
[EXTERNAL_DOWNLOADS]: Contains multiple references and links to external security tools and research repositories on GitHub and other developer platforms, such as GraphQLmap, InQL, and GQLSpection.
[INDIRECT_PROMPT_INJECTION]: The skill provides a large volume of attack payloads that an agent may be instructed to use, establishing a vulnerability surface when the agent processes response data from external targets.
Ingestion points: The skill contains a reference library of GraphQL injection payloads within SKILL.md.
Boundary markers: Includes a 'Legal Notice' and 'Authorized Use Only' header to provide operational context for the agent.
Capability inventory: The skill is designed for use with REA MCP tools for identifying injection sinks and applying payloads via function analysis.
Sanitization: The instructions do not define sanitization procedures for handling data received from target endpoints after injection attempts.

payloads-graphql