mph-substack-writer
Pass
Audited by Gen Agent Trust Hub on Sep 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from an external file to define its linguistic patterns, which could potentially contain malicious instructions.
- Ingestion points: The
SKILL.mdfile instructs the agent to read../mph-voice-refresh/VOICE.mdto identify phrasing patterns and retired impersonations. - Boundary markers: The instructions do not specify any delimiters or boundary markers to separate the external content from the skill's core instructions, nor do they include warnings for the agent to ignore any embedded directives within the voice file.
- Capability inventory: The skill possesses capabilities for text generation (Substack articles) and calls an external tool/skill named
mph-figureto generate images from data. - Sanitization: No sanitization, filtering, or validation steps are defined to process the content of
VOICE.mdbefore it enters the agent's context.
Audit Metadata