notion-policy-guardrails
Pass
Audited by Gen Agent Trust Hub on Sep 12, 2026
Risk Level: SAFECREDENTIALS_UNSAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [CREDENTIALS_UNSAFE]: The skill is designed to manage and audit Notion integration tokens. It provides scripts to detect hardcoded secrets (ntn_, secret_) in source files and ensures .env files are not committed to version control. This behavior supports secure credential management.
- [COMMAND_EXECUTION]: Uses shell commands (grep, git ls-files, npx ts-node) for security auditing and running implementation scripts. These tools are used within the context of the repository to enforce governance policies.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from Notion page titles and database properties when performing workspace audits.
- Ingestion points: Data is retrieved from the Notion API in references/implementation.md via notion.search and notion.databases.query.
- Boundary markers: No explicit delimiters are used for the retrieved data.
- Capability inventory: The skill has access to the Bash tool and can write/edit files as defined in SKILL.md.
- Sanitization: The retrieved data is printed for auditing or checked against regex patterns, but not sanitized before display.
- [DYNAMIC_EXECUTION]: Includes a script that generates a GitHub Actions workflow configuration. The generated content is based on static templates and is intended for manual integration into a CI/CD pipeline.
Audit Metadata