writing-fragments
Pass
Audited by Gen Agent Trust Hub on Jul 16, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection due to its core loop of reading and writing user-provided data.
- Ingestion points: The agent ingests untrusted data from the user conversation and subsequently re-reads this data from a markdown file on disk (SKILL.md).
- Boundary markers: While the skill uses horizontal rules (
---) to delimit fragments, it lacks explicit instructions or system-level delimiters to prevent the agent from executing instructions that may be contained within the fragments themselves. - Capability inventory: The skill is designed to perform file system read and write operations at user-specified paths.
- Sanitization: The skill does not implement sanitization, filtering, or validation of the text read from the markdown file before it is re-introduced into the agent's context.
Audit Metadata