art-style
Warn
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: MEDIUMCREDENTIALS_UNSAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- CREDENTIALS_UNSAFE (MEDIUM): The script
tools/run-generate.shutilizes the--env-file=../.envflag when callingtsx. This method for loading environment variables is risky as it can expose sensitive API keys or system secrets to the execution context of the agent. - COMMAND_EXECUTION (MEDIUM): The script
tools/run-generate.shtraverses five levels up the directory tree (../../../../..) to locate a server directory. This behavior breaks the expected encapsulation of an agent skill and allows the skill to interact with the broader host file system. - PROMPT_INJECTION (LOW): The skill is susceptible to indirect prompt injection because it incorporates external data from
/storage/hooks/into its workflow without sanitization. - Ingestion points: Data is read from
/storage/hooks/{brand}-{date}.md. - Boundary markers: No delimiters or instructions are provided to the agent to treat this file content as untrusted data.
- Capability inventory: The skill executes shell scripts and writes output to the file system.
- Sanitization: There is no evidence of validation or filtering for the content retrieved from the hook files.
Audit Metadata