speckit-git-feature
Pass
Audited by Gen Agent Trust Hub on Oct 12, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user input and interpolates it into shell commands and environment variables.\n
- Ingestion points:
$ARGUMENTSand user-providedGIT_BRANCH_NAMEinSKILL.md.\n - Boundary markers: None present to delimit user content or instruct the agent to ignore embedded commands.\n
- Capability inventory: Spawns shell processes (
bashandpowershell) to execute scripts located in the.specify/directory.\n - Sanitization: The instructions do not specify any sanitization, escaping, or validation logic for user-provided strings before they are used in command-line arguments.\n- [COMMAND_EXECUTION]: The skill invokes local scripts with arguments and environment variables derived from user input. This pattern creates a command injection risk if the input contains shell metacharacters (e.g., backticks, subshells, or command separators) that escape the surrounding double quotes used in the script invocation.\n- [DYNAMIC_EXECUTION]: The skill executes scripts from project-relative paths (e.g.,
.specify/extensions/git/scripts/). This behavior relies on the integrity of the project workspace; if the workspace is untrusted, these scripts could be replaced with malicious code which the agent would then execute.
Audit Metadata