Dex CRM

AdvisoryAudited by Static analysis on Apr 30, 2026.

Overview

No suspicious patterns detected.

Findings (0)

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

Anyone or any agent workflow using this key can access the Dex account according to the key’s permissions.

Why it was flagged

The skill needs a Dex API key that grants delegated access to the user’s CRM account. This is expected for the integration, but it is sensitive account authority.

Skill content
Set `DEX_API_KEY` in gateway config env vars.
Recommendation

Use a dedicated or revocable Dex API key if available, store it only in the intended gateway config, and remove or rotate it when no longer needed.

What this means

Mistaken or over-broad instructions could change or delete contacts, notes, or reminders in Dex.

Why it was flagged

The skill exposes write and delete CRM operations. The confirmation instruction is a good control, but these actions can permanently affect account data.

Skill content
Always confirm before creating, updating, or deleting contacts/notes/reminders
Recommendation

Confirm record IDs and intended changes before allowing create, update, or delete operations, especially for deletes or bulk edits.

What this means

A false positive in the junk-contact heuristic could archive real contacts in Dex.

Why it was flagged

The included cleanup script can bulk archive contacts selected by heuristics. README.md documents a dry-run mode, but a normal run performs account mutations.

Skill content
api_put(f"/contacts/{j['id']}", {
                "changes": {"is_archived": True}
            })
Recommendation

Run `python3 scripts/dex-cleanup.py --dry-run` first, review the listed contacts, and only run the non-dry-run command if the results are acceptable.

What this means

Personal contact information may be shown to or processed by the agent during CRM tasks.

Why it was flagged

The skill may retrieve personal CRM details and notes into the agent context. This is expected for a CRM skill, but the data can be sensitive.

Skill content
Look up contact details (phone, email, birthday)
Recommendation

Use the skill only in trusted conversations and avoid asking it to retrieve more contact data than needed.

What this means

Users have less external context for verifying authorship or updates.

Why it was flagged

The registry does not provide upstream provenance. No risky install step or hidden dependency is shown, so this is a provenance notice rather than a behavioral concern.

Skill content
Source: unknown
Homepage: none
Recommendation

Review the included SKILL.md and script before use, and prefer installing from a trusted publisher or verified source when available.