speckit-specify
Pass
Audited by Gen Agent Trust Hub on Oct 12, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill implements a pre-execution and post-execution hook system by reading
.specify/extensions.yml. It identifies commands to run and triggers them using theEXECUTE_COMMANDinstruction, which allows local project configuration to control the execution of arbitrary slash commands. - [INDIRECT_PROMPT_INJECTION]: The skill processes natural language descriptions from the user to generate structured specifications, creating a surface for instructions embedded in data to override intended behavior.
- Ingestion points: User feature descriptions provided as
$ARGUMENTSinSKILL.md. - Boundary markers: Absent. User input is directly integrated into the prompt context without delimiters or 'ignore' instructions.
- Capability inventory: Includes directory creation (
mkdir), file writing (spec.md,feature.json, and checklists), and command execution (EXECUTE_COMMAND). - Sanitization: Absent. The skill lacks validation or escaping for the user-provided text, relying on the AI to 'extract key concepts' directly from the input.
Audit Metadata