pentesting-web-apps
Pass
Audited by Gen Agent Trust Hub on Oct 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The helper scripts utilize
bash -lcto execute external security tools, including curl, sqlmap, and OWASP ZAP, based on the user-provided target URL. The scripts employprintf %qfor shell escaping to mitigate basic command injection risks, but the execution model relies on the availability and behavior of these external binaries on the host system. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze content from external, untrusted web applications. This creates a surface for indirect prompt injection where a target site could host malicious text intended to subvert the agent's instructions during the analysis phase.
- Ingestion points: Data is ingested via web crawling (playwright-crawl.mjs), header inspection (auth-analysis.sh), and directory discovery (crawl-surface.sh).
- Boundary markers: The instructions do not define clear delimiters or "ignore instructions" wrappers for data returned from the target URL.
- Capability inventory: The skill has capabilities for file system writing (results logs), network access (probing target URLs), and execution of specialized security tools (sqlmap, dalfox, zap.sh).
- Sanitization: The
normalize_urlfunction incommon.shvalidates the URL scheme and format, andlogic-tester.shperforms basic sanitization on cookie headers to prevent newline injection.
Audit Metadata