yetong-fitness-xhs-topic
Pass
Audited by Gen Agent Trust Hub on Oct 3, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute a bundled Python script (
scripts/check_profile.py) with a user-provided file path as an argument. While this is used for validation, it requires the agent platform to properly handle shell argument escaping to prevent command injection. - [INDIRECT_PROMPT_INJECTION]: The skill reads and processes the content of an external Markdown file (
person-dna.md) to personalize its output. This creates a surface where instructions embedded in the user-provided file could potentially override agent behavior. - Ingestion points: The content of the file specified at the path provided by the user is read into the agent's context.
- Boundary markers: The skill includes a detailed 'Profile Contract' (
references/profile-contract.md) that defines how the agent should interpret different data fields and usage permissions. - Capability inventory: The skill executes a local Python script and generates content based on the ingested profile data.
- Sanitization: The skill uses a script to validate the file structure and frontmatter, but it does not perform semantic sanitization of the text content to filter out potential instructions.
- [DYNAMIC_EXECUTION]: The skill performs runtime validation of profile files using a local Python script, which is a dynamic method for ensuring data integrity before processing.
Audit Metadata