fable-native-code
Pass
Audited by Gen Agent Trust Hub on Oct 10, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided file content (
target_file) to match coding idioms. This creates a surface where instructions embedded in the codebase could influence the agent's behavior during the modification process. - Ingestion points: The
target_fileinput inSKILL.mdis used to absorb file conventions. - Boundary markers: No explicit markers are defined to distinguish code from potential embedded natural language instructions in the target file.
- Capability inventory: The skill possesses workspace mutation capabilities (
mutatesWorkspace: true) and uses tools likereplace_file_contentto modify the source code. - Sanitization: There is no evidence of sanitization or filtering of the input file content to prevent instruction injection.
Audit Metadata