skills/mukul975/anthropic-cybersecurity-skills/performing-graphql-introspection-attack/Gen Agent Trust Hub
performing-graphql-introspection-attack
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill contains Python implementation for active GraphQL attack techniques, including generating Denial-of-Service (DoS) payloads via deeply nested queries, wide queries (field duplication), and circular fragments.\n- [COMMAND_EXECUTION]: Includes scripts for alias-based batching attacks, which can be used to bypass per-request rate limiting during credential brute-forcing.\n- [INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection through the ingestion of external data.\n
- Ingestion points: Processing of external API responses (JSON schema data and error messages) from user-specified targets in SKILL.md and scripts/agent.py.\n
- Boundary markers: No boundary markers or instructions to ignore embedded content are present.\n
- Capability inventory: The skill has the capability to perform network requests (requests.post) and write to the local file system (graphql_schema.json).\n
- Sanitization: No sanitization or validation of the external API responses is performed before outputting data to the console.
Audit Metadata