openspec-config
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill manages a configuration file (openspec/config.yaml) containing fields (context, rules) that are explicitly designed to be injected into AI instructions for artifacts.
- Ingestion points: The context and rules fields in openspec/config.yaml.
- Boundary markers: The skill specifies that content is delimited using and tags, which helps mitigate accidental instruction following.
- Capability inventory: The skill utilizes the openspec CLI to manage settings but does not provide local scripts or execute remote code.
- Sanitization: No sanitization or validation of the injected content is mentioned, meaning the data in the config file could potentially contain malicious instructions.
Audit Metadata