skill-family-care
Pass
Audited by Gen Agent Trust Hub on Nov 13, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill requires the execution of shell commands to run local Python scripts (
inventory_skills.pyandinject_family_header.py). These scripts are used to gather skill inventory and modify the headers of other local skill files. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from local sources to influence its actions and the content it writes to other files.
- Ingestion points: The skill reads from
<USER_HOME>\OneDrive\.USR\SKILL-MAP.mdand~/.claude/skills/skill-explorer/config.jsonto determine family assignments and routing information. - Boundary markers: The instructions do not specify any delimiters or boundary markers for the data ingested from the mapping files.
- Capability inventory: The skill possesses the ability to execute shell commands (via Python) and write to other skill files located in the user's skill directory.
- Sanitization: While the instructions mention that scripts have 'safety gates' to only modify editable/user-owned skills, there is no evidence of sanitization or validation of the content being injected into those files from the source mapping documents.
Audit Metadata