openspec-continue-change
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes local
openspecCLI commands such aslist,status, andinstructionsto interact with the project state. These commands are fundamental to the skill's intended purpose and are executed with parameters derived from the user's selection. - [INDIRECT_PROMPT_INJECTION]: The skill processes external data from the CLI to generate project artifacts. 1. Ingestion points: Data enters the context via
openspec status --jsonandopenspec instructions --jsonoutputs. 2. Boundary markers: None explicitly defined for the CLI JSON output; the agent parses fields directly. 3. Capability inventory: The skill has the capability to write files to the local file system (outputPath). 4. Sanitization: The skill uses atemplatestructure to constrain the output, which helps ensure that instructions from the CLI guide the content generation without introducing arbitrary code or malicious commands.
Audit Metadata