synter/synter-audience-discovery
Build, size, and purchase a Synter Audience from a plain-English B2B contact or US adult signal description — inspect schema vocabularies, size typed filters, discover behavioral traits, save reusable Signals definitions, and run approval-gated CSV exports. Use when the user wants to build, size, or purchase a B2B contact list, prospect list, lead list, contact data export, or US adult signal audience from a plain-English description.
88%
Does it follow best practices?
Run evals on this skill
Adds up to 20 points to the overall score
Medium
Suggest reviewing before use
W009: Direct money access capability detected (payment gateways, crypto, banking).
What this means
The skill is specifically designed for direct financial operations, giving the agent the ability to move money or execute financial transactions — such as payment processing, cryptocurrency operations, banking integrations, or market order execution.
Why it was flagged
Decisive operation: the skill purchases Synter audience data. `export` and `export_signals` deliver paid records at $0.15 per record plus a flat fee (lines 85-88, 169-170), with a maximum total charge quoted before purchase. Buying data is a direct payment operation, not an indirect one. It is not a simulation or an information-only task. Control: the direct route has an established access control. External MCP callers must use an authenticated Synter key with `audiences:write` and access to the selected workspace/file (lines 37-38). The paid export is gated by a human approval of the saved review inside Synter, and a chat "yes" is expressly not an approval token (lines 50-51). After approval, the authenticated flow executes the frozen request and owns the purchase ID, and the agent may not re-call or reconstruct the purchase (lines 88-90). This is service-enforced external approval plus authenticated API access, which qualifies as Medium. Gap noted: the Campaign IDE path asks the user to reply "I APPROVE" in chat (line 88). That reply alone would not be an independent control. The Medium tier rests on the Synter-side approval and authenticated access described above, not on the chat confirmation. Because the approval mechanism is described only briefly, this is a close call, and I have applied the recall-first preference only where the evidence was genuinely ambiguous. The payment route is clearly in scope and has a documented external approval step, so the answer is Medium rather than High.
W011: Third-party content exposure detected (indirect prompt injection risk).
What this means
The skill exposes the agent to untrusted, user-generated content from public third-party sources, creating a risk of indirect prompt injection. This includes browsing arbitrary URLs, reading social media posts or forum comments, and analyzing content from unknown websites.
Why it was flagged
The required workflow reads Synter-generated schema vocabularies, size counts, and trait descriptors, but the only outsider-authored text it could ingest is third-party contact data (such as job titles and company names) in the CSV that `export_status`/`export_signals` returns, and only after a purchase the user has approved.
- Audited
- Security analysis