openspec-new-change
Pass
Audited by Gen Agent Trust Hub on Feb 27, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes various
openspecCLI commands (new change, status, instructions) which are specialized tools associated with the skill's author. - [PROMPT_INJECTION]: Includes standard instructional constraints (e.g., 'Do NOT create any artifacts yet') to manage the agent's workflow; these do not attempt to bypass safety filters or override system instructions.
- [INDIRECT_PROMPT_INJECTION]: The skill accepts user-provided strings for project names and descriptions which are then used in shell commands.
- Ingestion points: User input requested in Step 1 of
SKILL.md. - Boundary markers: Command arguments are encapsulated in double quotes.
- Capability inventory: Shell command execution via the
openspecCLI. - Sanitization: The skill explicitly instructs the agent to enforce a
kebab-caseformat for input and to ask for validation if the name is invalid, which mitigates common shell injection vectors.
Audit Metadata