speckit-baseline
Pass
Audited by Gen Agent Trust Hub on Jun 23, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes a local repository script at
.specify/scripts/bash/create-new-feature.shusing parameters derived from the code it analyzes as part of Workflow Step 5. - [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it reads and processes external source code that could contain adversarial instructions.
- Ingestion points: Workflow Step 2 reads arbitrary file contents and expands globs within the repository.
- Boundary markers: No explicit delimiters or instructions are provided to separate the analyzed source code from the agent's internal operational instructions.
- Capability inventory: The skill enables the agent to execute shell scripts and write multiple files to the local file system.
- Sanitization: Workflow Step 5 provides specific guidance on escaping single quotes in shell arguments to minimize command injection risk.
Audit Metadata