define-my-content-pillars
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests and analyzes 10 to 15 social media posts provided by the user or fetched via an external API. This represents an attack surface for indirect prompt injection, where malicious instructions could be embedded in the posts to influence the agent's behavior.
- Ingestion points: User-pasted posts in SKILL.md Step 1 and posts retrieved via VoiceMoat API tools (list_posts, get_post) described in the connection section.
- Boundary markers: No specific delimiters or "ignore embedded instructions" warnings are implemented in the clustering or testing logic.
- Capability inventory: The skill reads post text and account data; it does not have the capability to write to the file system or execute shell commands.
- Sanitization: No content filtering or sanitization of the social media posts is described.
- [DATA_EXFILTRATION]: The skill describes an optional integration where tool calls (list_posts, get_analytics, get_voice_insights) result in data being read from or sent to 'voicemoat.com'. As this domain is not on the standard whitelist for network operations, it is flagged as a data transmission vector, although the skill documentation notes that it uses OAuth and no API keys for this purpose.
Audit Metadata