skills/zixun-github/ai-dlc/spec-init/Gen Agent Trust Hub

spec-init

Warn

Audited by Gen Agent Trust Hub on Jul 28, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses powershell -NoProfile -ExecutionPolicy Bypass -File to execute a local script. This pattern is often used to circumvent local security configurations designed to prevent script execution.
  • [COMMAND_EXECUTION]: The spec-create-branch.sh and spec-create-branch.ps1 scripts execute various git commands (checkout, branch, fetch) and file system operations based on user-provided inputs (--short-name, --source-file). While the scripts attempt to validate some inputs (e.g., kebab-case for short names), the interaction between user-controlled strings and shell execution poses a risk of command injection if validation is bypassed.
  • [DATA_EXFILTRATION]: The skill's documented behavior includes an explicit step to delete the source requirement file after processing (Remove-SourceFile in PowerShell, rm -f in Bash). While this is disclosed to the user, such automated deletion of local files can be used by malicious skills to destroy evidence or remove sensitive data after exfiltration (though no explicit exfiltration was found in the static analysis).
Audit Metadata
Risk Level
MEDIUM
Analyzed
Jul 28, 2026, 08:08 AM
Security Audit — agent-trust-hub — spec-init