prompt-atom-formalization
Pass
Audited by Gen Agent Trust Hub on Jun 16, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill follows established prompt engineering governance patterns (RAW -> CLEANED -> ELEVATED) to organize developer intent.
- [COMMAND_EXECUTION]: While the skill body mentions command-line tools like
chezmoi, these are used solely as illustrative examples within prompt atom text (e.g., demonstrating a 'constraint') and are not intended for execution by the agent. - [DATA_EXPOSURE]: The skill declares
reads-filesystemandcreates-filesin its metadata. This is consistent with its stated purpose of reading source text and appending formalized atoms to a local registry file. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted input ('source text', 'session exports'). It mitigates risks by requiring the agent to classify and 'formalize' (normalize) the intent into a structured YAML schema rather than directly executing the content.
Audit Metadata