openspec-continue-change
Pass
Audited by Gen Agent Trust Hub on Apr 22, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill invokes the
openspecCLI tool with arguments derived from previous command outputs and user input to manage changes and artifacts.- [PROMPT_INJECTION]: The skill contains an indirect prompt injection surface as it consumes external instructions to guide its behavior and file operations. - Ingestion points: The agent reads instructions, project context, rules, and file paths from the JSON output of the
openspec instructionscommand and completed dependency files. - Boundary markers: No delimiters or specific instructions to ignore embedded prompts are implemented for the data ingested from the CLI.
- Capability inventory: The skill is capable of executing shell commands via the
openspecutility and writing files to paths specified by the tool's output. - Sanitization: The skill lacks mechanisms to validate the
outputPathor sanitize the instructional content provided by the external tool before execution.
Audit Metadata