write-a-prompt
Pass
Audited by Gen Agent Trust Hub on Aug 26, 2026
Risk Level: SAFE
Full Analysis
- [DATA_EXPOSURE_AND_EXFILTRATION]: The skill instructions proactively address data safety by mandating that "raw secrets" be omitted from extracted context and that the final quality check ensures no "exposed secrets" are contained in the output.
- [INDIRECT_PROMPT_INJECTION]: The skill processes external data including user requests, notes, and conversation history, which creates an ingestion surface.
- Ingestion points: User intent, source material, and conversation context as defined in SKILL.md.
- Boundary markers: Explicit boundary markers or delimiters for input context are absent.
- Capability inventory: No dangerous capabilities such as subprocess calls, file writing, or unauthorized network operations are present across the skill.
- Sanitization: The skill instructions require the agent to sanitize output by omitting secrets, obsolete directions, and irrelevant dialogue.
- [UNVERIFIABLE_DEPENDENCIES_AND_REMOTE_CODE_EXECUTION]: The skill refers to the official OpenAI prompting guide to inform its logic. This reference targets a well-known service for documentation purposes and does not involve downloading or executing external scripts.
Audit Metadata