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 fetches external JSON data from openspec CLI commands to provide runtime context and operation guidance. Ingestion points: Data is ingested through shell command outputs (openspec instructions archive/specs) in SKILL.md (Steps 1 and 4). Boundary markers: The skill explicitly instructs the agent to maintain separation between external guidance and core workflow steps and to prioritize controlling inputs over conflicting external advice. Capability inventory: The skill uses the Bash tool to execute CLI commands (openspec) and perform file system operations (mkdir, mv). Sanitization: Instructions include guardrails that forbid the agent from using external data to infer file paths, skip confirmation prompts, or modify command contracts.
- [COMMAND_EXECUTION]: The skill performs directory management using mkdir and mv commands. These operations utilize variables such as changeRoot and planningHome.changesDir which are derived from the output of previous CLI commands, creating a dependency on the integrity of the openspec tool output.
Audit Metadata