offensive-file-upload
Pass
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill defines a workflow for auditing external source code, binaries, and web bundles. This establishes a surface where malicious instructions embedded in the data being audited could attempt to influence the agent's logic or bypass its instructions.
- Ingestion points: The agent is instructed to inspect source code, binaries, and configuration files (SKILL.md).
- Boundary markers: No explicit boundary markers or delimiters are defined to isolate the audited content from the agent's instruction set.
- Capability inventory: The skill workflow involves using analysis tools (rg, AST analysis), disassembly tools, and potentially executing local test suites (npm test).
- Sanitization: The instructions do not include steps for sanitizing or escaping audited content before processing.
- [NO_CODE]: The skill contains only documentation, workflow descriptions, and configuration metadata. There are no scripts, binaries, or automated tasks included that would execute without further agent planning or external tool invocation.
Audit Metadata