openspec-propose
Pass
Audited by Gen Agent Trust Hub on Sep 22, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill instructions and logic follow safe practices for planning workflows. No malicious patterns such as exfiltration, persistence, or obfuscation were detected.
- [COMMAND_EXECUTION]: The skill relies on the
openspecCLI tool to perform its operations. This activity is restricted to the specific CLI binary allowed by the platform configuration. - [INDIRECT_PROMPT_INJECTION]: The skill ingests external data from the
openspecCLI and user input to generate planning artifacts. 1. Ingestion points:openspec instructionsJSON output, user descriptions, and project dependency files (e.g.,proposal.md,specs/). 2. Boundary markers: The skill defines a strict 'Planning boundary' and instructs the agent to stop after planning, explicitly forbidding implementation or source code edits. 3. Capability inventory: UsesBash(openspec:*)for status checks and artifact generation, and writes files toresolvedOutputPath. 4. Sanitization: The skill guides the agent to follow specific templates and schemas provided by the CLI, reducing the risk of unauthorized command execution through processed data.
Audit Metadata