re0-order
Pass
Audited by Gen Agent Trust Hub on Nov 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes listings (lists, tables, sets) which may contain untrusted content designed to override agent instructions. The workflow additionally includes a repository-wide search for mirrored items, which could expose the agent to malicious content in other files.
- Ingestion points: Workflow Step 1 (identifying the listing) and Workflow Step 6 (searching for mirrored copies across the repository based on item content).
- Boundary markers: Absent; the skill does not define delimiters or instruct the agent to ignore instructions embedded within the listings it processes.
- Capability inventory: The skill uses file system reading/writing and explicitly recommends the use of scripts for structural moves.
- Sanitization: None; the provided verification script checks for multiset equality but does not sanitize or validate the content of the items.
- [DYNAMIC_EXECUTION]: The skill's instructions suggest that the agent should 'script a large structural move' rather than performing it manually (Rules). This dynamic generation and execution of code based on untrusted listing content introduces a risk of code injection if the listing contains characters designed to break out of the script's syntax.
Audit Metadata