offensive-api-abuse
Pass
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: SAFENO_CODEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill provides a structured framework for API business logic and automated abuse resilience assessment. The content is purely instructional, outlining steps for discovery, verification, and triage without including any executable code or configuration that would perform automated actions without user interaction.\n- [INDIRECT_PROMPT_INJECTION]: The skill's primary function involves the agent ingesting and analyzing external source code, binaries, and API bundles during the 'Surface & Architecture Inventory' phase. This exposes the agent to instructions that might be embedded in the audited data, which is a known risk for any security analysis tool. However, the skill does not contain instructions to execute these potentially malicious patterns, only to analyze them.\n
- Ingestion points: Target source code, binaries, and web bundles identified during inventory (SKILL.md, Operational Workflow step 1).\n
- Boundary markers: The instructions do not define specific delimiters or isolation protocols for processing the untrusted content.\n
- Capability inventory: Mentions the use of 'REA MCP tools', search utilities like 'rg', AST analysis, decompilation tools, and executing 'npm test' for verification.\n
- Sanitization: No specific sanitization or filtering instructions for the external content are provided in the workflow.
Audit Metadata