github-search
Fail
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: HIGHCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [Indirect Prompt Injection] (HIGH): The skill ingests untrusted data from GitHub search results which is then processed by the agent. \n
- Ingestion points: Search results from
gh searchandgh apicalls in SKILL.md. \n - Boundary markers: None provided to isolate external data from the agent's instruction context. \n
- Capability inventory: The agent has access to the
Bashtool, enabling full command execution on the host. \n - Sanitization: No sanitization or safety filtering is applied to the retrieved GitHub content. \n- [Dynamic Execution] (MEDIUM): The skill uses command substitution
$(...)andjqto dynamically construct shell commands for API requests, which could be exploited for shell injection if the agent does not properly escape user-provided query strings.
Recommendations
- AI detected serious security threats
Audit Metadata