security-nextjs
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides specific shell commands (grep, ripgrep, fd) and a custom bash script (scripts/scan.sh) to perform automated security scans on the local filesystem.- [DATA_EXFILTRATION]: The auditing process and the provided scanner script specifically target and read sensitive files, such as environment variable files (.env*), which are likely to contain private credentials and API keys. While the findings are reported locally, this exposes sensitive data to the agent's context.- [DYNAMIC_EXECUTION]: The skill includes a standalone shell script (scripts/scan.sh) that the agent is expected to execute to automate the auditing process across the project directory.- [INDIRECT_PROMPT_INJECTION]: The skill's primary function involves reading and processing untrusted content from the codebase being audited, creating a vulnerability to embedded instructions.
- Ingestion points: The script and commands read from .env files, next.config.js, middleware configurations, and various application source code files.
- Boundary markers: There are no explicit instructions or delimiters used to separate the analyzed file content from the agent's internal reasoning logic.
- Capability inventory: The agent has the capability to execute shell commands and scripts to process the data found in these files.
- Sanitization: The skill does not implement any filtering or sanitization of the content extracted from project files before it is processed by the agent.
Audit Metadata