launchdarkly-flag-drift
Pass
Audited by Gen Agent Trust Hub on Oct 12, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external sources, specifically LaunchDarkly flag configurations and repository content, which are used to drive search operations and populate pull request descriptions.
- Ingestion points: Data is pulled from the LaunchDarkly API (via
get-flagandlist-flagstools) and local source files. - Boundary markers: The instructions do not specify the use of delimiters or specific warnings to ignore embedded instructions in the ingested data.
- Capability inventory: The agent is authorized to perform file system searches (using
rg), modify source code, execute local build/test scripts, and perform Git operations. - Sanitization: There are no explicit instructions to sanitize flag keys or variation values before they are used in shell-based search commands or automated PR descriptions.
- [COMMAND_EXECUTION]: The skill requires the agent to execute project-specific validation commands (such as
npm test,make, or linting tools) discovered within the repository's configuration files. This execution is a necessary part of the intended workflow to ensure code quality before submitting changes.
Audit Metadata