Sophiie AI Office Manager

PassAudited by ClawScan on May 10, 2026.

Overview

This appears to be a coherent Sophiie API wrapper, but it can use your Sophiie API key to read and change CRM data and send SMS/calls, so impactful actions should be confirmed.

This skill looks purpose-aligned for managing Sophiie, but it is not risk-free: it can access customer/business data and perform real actions such as updating records, policies, FAQs, SMS, and calls. Verify the publisher/source, use the least-privileged or test API key available, and require confirmation before any outbound communication, deletion, or customer-facing change.

Findings (4)

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

What this means

The agent could change CRM records or contact customers if instructed to do so.

Why it was flagged

The skill can perform external side-effecting actions such as sending calls/SMS and deleting or modifying Sophiie records. This is disclosed and aligned with the office-manager purpose, but it is high-impact.

Skill content
POST `https://api.sophiie.ai/v1/calls` ... POST `https://api.sophiie.ai/v1/sms` ... DELETE `https://api.sophiie.ai/v1/leads/{id}`
Recommendation

Use this skill only when you intend to manage Sophiie data, and require explicit confirmation before sending messages/calls, deleting records, or changing customer-facing FAQs/policies.

What this means

Anyone or any agent run with this environment variable can act through the Sophiie API within the key's permissions.

Why it was flagged

The skill uses a bearer API key for the user's Sophiie account. This is expected for the integration, but the credential controls access to business/customer data.

Skill content
All requests use `Authorization: Bearer <key>` where the key is `SOPHIIE_API_KEY`.
Recommendation

Prefer a scoped or sandbox/test key where possible, rotate keys if exposed, and avoid enabling this skill in untrusted sessions.

What this means

Customer messages or transcripts could include sensitive information or prompt-like text that may influence the agent if not handled as data.

Why it was flagged

Inquiry data may include customer-supplied text that enters the agent context. Such text is useful business data but should not be treated as instructions to the agent.

Skill content
Returns the inquiry with expanded source data (call transcripts, SMS messages, webform submissions, etc.)
Recommendation

Treat retrieved inquiries, transcripts, SMS, emails, and webform text as untrusted content; do not let them override the user's instructions.

What this means

A user has less assurance that the skill came from the official Sophiie publisher.

Why it was flagged

The package provenance is not identified in the supplied metadata. No hidden installer or remote dependency is shown, but provenance matters because the skill handles a live API key.

Skill content
Source: unknown
Recommendation

Verify the skill against Sophiie's official documentation or publisher before using a production `sk_live_*` API key.