openspec-archive-change
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external CLI tool outputs and local markdown files which are used to influence the agent's behavior.
- Ingestion points: The skill reads JSON outputs from
openspec instructions archive,openspec status, andopenspec instructions specs, as well as the contents of thetasks.mdfile. - Boundary markers: No explicit delimiters are specified for the ingested content, though the skill treats them as structured data or advisory inputs.
- Capability inventory: The agent possesses the capability to perform file system operations (
mkdir,mv) and execute further shell commands via theBashtool. - Sanitization: The instructions include a safeguard requiring the agent to "report the conflict and preserve the controlling value" if external context or guidance conflicts with the core workflow or user intent.
Audit Metadata