exit-code-fast-checking
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill contains an 'Agent Operational Directive' section that uses imperative language ('MANDATORY') to instruct the agent to modify its default behavior regarding command output verbosity.\n- [COMMAND_EXECUTION]: The skill provides a Python implementation of
execute_command_with_exit_guardand shell snippets that facilitate the execution of arbitrary CLI tools usingsubprocess.runand redirection patterns.\n- [INDIRECT_PROMPT_INJECTION]: The skill defines a process for capturing and displaying output from external tools, which creates a surface for indirect instructions if the tools are compromised or manipulated.\n - Ingestion points: External command output is captured via
subprocess.runand extracted fromresult.stderrorresult.stdoutinSKILL.md.\n - Boundary markers: None identified; command output is directly concatenated into the confirmation message.\n
- Capability inventory: The skill includes functions designed to wrap and execute terminal commands through
subprocess.run.\n - Sanitization: None; while the output is truncated to the first 25 lines, there is no escaping or filtering of the content before it is returned to the agent context.
Audit Metadata