verify
Pass
Audited by Gen Agent Trust Hub on Nov 16, 2026
Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes specific management and checking scripts located within the local environment at
~/.claude/skills/gstack/bin/. These scripts handle initialization, telemetry, and test execution. - [DYNAMIC_EXECUTION]: The skill is designed to receive and execute dynamic instruction blocks (
GSTACK_INSTRUCTION_BEGIN) provided by thegstack-skill-startutility at runtime. - [INDIRECT_PROMPT_INJECTION]: The skill processes instructions from external tool outputs. To mitigate potential injection, it implements security boundaries, such as requiring instructions to originate only from the designated start command and requiring a matching
SESSION_IDto be considered valid. - [DATA_EXFILTRATION]: Session outcome telemetry and durable learnings are captured and sent to local scripts for logging and analysis.
- [SAFE]: The skill includes explicit security ground rules that prevent the agent from displaying environment variables, process tables, or git configurations that might contain sensitive access tokens.
Audit Metadata