openspec-archive-change
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external data from
openspec instructionscommands which can includecontextandoperationGuidancethat the agent is instructed to follow. - Ingestion points: Data is ingested in Step 1 (archive instructions) and Step 4 (spec instructions) via the
openspecCLI. - Boundary markers: The skill explicitly instructs the agent to treat this data as advisory/optional and prevents it from overriding explicit user choices, CLI contracts, or resolved paths.
- Capability inventory: The skill has the ability to execute
openspecCLI commands, create directories (mkdir), and move files/directories (mv). - Sanitization: There is no explicit content sanitization mentioned for the fetched instructions other than ensuring valid JSON format.
- [COMMAND_EXECUTION]: The skill constructs and executes shell commands (
openspec,mkdir,mv) using variables derived from the local environment and user input. - Evidence: Uses variables like
"<name>","<changeRoot>", and"<planningHome.changesDir>"within bash blocks inSKILL.mdto automate archive operations.
Audit Metadata