openspec
Pass
Audited by Gen Agent Trust Hub on Sep 29, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes the
openspecCLI tool to perform project-level actions. These include initializing the environment, checking change status, and validating artifact formats. - [EXTERNAL_DOWNLOADS]: The skill references the installation and execution of the
@fission-ai/openspecpackage vianpxorpnpm. This is the primary tool for the specification-driven workflow and is fetched from the official NPM registry. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to process and act upon content within project-specific specification and design documents, which could potentially contain untrusted instructions if files are manipulated externally.
- Ingestion points: Files located in
openspec/specs/andopenspec/changes/including proposals, technical designs, and task lists. - Boundary markers: Uses mandatory Markdown structures and keywords (MUST, SHALL, GIVEN, WHEN, THEN) to delimit requirements. Employs
openspec validate --strictto enforce these boundaries. - Capability inventory: The agent has the capability to modify project files and execute local CLI commands based on the content of these documents.
- Sanitization: Validates the schema and structure of artifacts before implementation, reducing the risk of accidental execution of malformed instructions.
Audit Metadata