openspec-archive-change
Pass
Audited by Gen Agent Trust Hub on Sep 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes data from external command outputs that may contain instructions meant to influence agent behavior. • Ingestion points: Reads JSON from
openspec instructions archiveandopenspec instructions specs(Step 1 and 4). • Boundary markers: Includes directives to treat external instructions as advisory and subordinate to built-in protocols and user inputs. • Capability inventory: Ability to create directories, move files, and executeopenspecCLI commands. • Sanitization: Instructs the agent to validate external guidance against 'controlling inputs' and report any conflicts. - [COMMAND_EXECUTION]: Uses
Bashto runopenspecCLI tools and filesystem commands (mkdir,mv) to perform archiving operations.
Audit Metadata