speckit
Pass
Audited by Gen Agent Trust Hub on Oct 6, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes natural language descriptions from users via the
/speckit.specifycommand. This input is interpolated into markdown templates (e.g.,assets/templates/spec-template.md) to create feature specifications. While this creates an attack surface for indirect prompt injection, it is the primary function of the skill and is managed by the underlying model's safety protocols and the skill's own instructions. - Ingestion points: User input enters via the
/speckit.specifycommand and is placed inspec.mdusing the template inassets/templates/spec-template.mdvia the$ARGUMENTSplaceholder. - Boundary markers: User input is wrapped in markdown formatting within generated files.
- Capability inventory: The skill performs file system operations (creating directories and markdown files) via the
scripts/init_spec_dir.pyscript. - Sanitization: No explicit sanitization or escaping of user input is performed by the provided scripts.
- [COMMAND_EXECUTION]: The skill includes Python scripts (
scripts/init_spec_dir.py,scripts/validate_spec.py) that perform local file system operations. These scripts are restricted to creating thespecs/directory structure and validating the presence of required markdown sections and files. They do not invoke shell commands or execute code from untrusted sources.
Audit Metadata