safe-settings
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external YAML configuration files to manage organization-level GitHub settings, which introduces a potential attack surface for indirect prompt injection if the configuration repository is compromised.
- Ingestion points: Configuration files located at
.github/settings.yml,.github/suborgs/*.yml, and.github/repos/*.yml(as described inSKILL.mdandreferences/schema.md). - Boundary markers: The skill relies on structured YAML schema for data parsing but does not include specific prompt-level delimiters or instructions to ignore embedded commands within the configuration values.
- Capability inventory: The skill's synchronization workflow executes
git clone,npm ci, andnpm run full-syncinassets/full-sync.yml, and utilizesghCLI commands for secret and variable management. - Sanitization: Input validation and sanitization are delegated to the underlying
safe-settingsProbot application logic; no additional sanitization is implemented within the skill's instructional protocol. - [REMOTE_CODE_EXECUTION]: The synchronization workflow is designed to download and execute code from a remote repository at runtime.
- Evidence: The GitHub Actions workflow in
assets/full-sync.ymlclones thegithub-community-projects/safe-settingsrepository, installs its dependencies vianpm ci, and executes the application logic usingnpm run full-sync. - [EXTERNAL_DOWNLOADS]: The skill fetches tool source code and uses external workflows from well-known repositories.
- Evidence:
assets/full-sync.ymlfetches code fromgithub.com/github-community-projects/safe-settingsand utilizes official GitHub Actions such asactions/checkoutandactions/setup-node, which are correctly pinned to specific commit hashes for security. - [COMMAND_EXECUTION]: The skill utilizes shell commands to manage repository state, credentials, and reconciliation processes.
- Evidence:
SKILL.mdinstructs the user to executegh variable setandgh secret setto configure GitHub App credentials, while the synchronization workflow inassets/full-sync.ymlexecutes shell-based installation and execution commands.
Audit Metadata