ai-native-backend-contract
Pass
Audited by Gen Agent Trust Hub on Jul 9, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill serves as a reference and schema definition for managing workflow state via GitHub-native features or markdown files. It does not contain any executable code, scripts, or network operations.
- [PROMPT_INJECTION]: No evidence of instructions designed to override safety filters, bypass guidelines, or extract system prompts was found.
- [CREDENTIALS_UNSAFE]: No hardcoded API keys, tokens, or credentials were detected. The mention of an API token in
backends/markdown-file-based.mdis a descriptive example of a blocker and not an actual secret. - [DATA_EXFILTRATION]: There are no network requests or patterns indicating the unauthorized transmission of data to external servers.
- [REMOTE_CODE_EXECUTION]: The skill does not perform any dynamic code execution, nor does it download or execute external scripts.
- [INDIRECT_PROMPT_INJECTION]: While the skill defines how to ingest untrusted data from GitHub issues or markdown files, it acts solely as a contract for data representation and does not perform the ingestion itself. The structure encourages the use of specific headers and YAML frontmatter which act as boundary markers for processed content.
Audit Metadata