add-git-tag
Pass
Audited by Gen Agent Trust Hub on May 14, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes the Bash tool to execute Git operations. In Step 4, it constructs a command
git -C "$REPO" tag -a "<version>" -m "<full_tag_message>"using strings directly gathered from the user. If the agent does not perform adequate escaping of shell-sensitive characters (such as backticks, semicolons, or dollar signs), it could lead to arbitrary command execution during the tag creation process. - [PROMPT_INJECTION]: The skill contains an indirect prompt injection surface because it processes untrusted user input that is then interpolated into a script for tool execution.
- Ingestion points: User input for tag version, achievements, and future plans in Step 1.
- Boundary markers: Step 2 provides a confirmation mechanism where the user reviews the formatted message before the command is executed, though the underlying shell command is not shown.
- Capability inventory: Shell execution via Bash for repository status checks, tagging, and pushing (Steps 3, 4, and 5).
- Sanitization: Absent. There are no instructions for the agent to validate or sanitize the user-provided strings to prevent shell escaping or injection.
Audit Metadata