setup-cracked
Pass
Audited by Gen Agent Trust Hub on Nov 10, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to verify the local environment by checking for the availability of high-privilege development tools such as
python3,g++,clang++,javac,node,go,sqlite3, and the GitHub CLI (gh). This involves the execution of shell commands to probe for binary existence and versions. - [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied data in the form of
target_companies, which is then passed to external tools like "company-intel". This creates a surface where an attacker could theoretically embed instructions within company names to manipulate the agent's logic during the intel-gathering phase. - Ingestion points: User-provided list of company names in
SKILL.md. - Boundary markers: Absent; there are no clear delimiters or "ignore previous instructions" warnings surrounding the interpolated company names.
- Capability inventory: The skill suite is verified to have execution capabilities for Python, C++, Java, Node.js, Go, and SQL, along with authenticated GitHub access.
- Sanitization: No sanitization or validation logic is present to filter malicious instructions from the input data.
Audit Metadata