Percept Speaker ID
PassAudited by ClawScan on May 1, 2026.
Overview
The skill is coherent and instruction-only, but it stores speaker identity information and uses it to decide who can issue voice commands.
This skill appears appropriate for speaker identification and voice-command authorization. Before installing, make sure you trust the connected Percept/Omi setup, restrict access to the speaker dashboard and registry file, and verify that only intended people are marked as owner or approved.
Findings (2)
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.
If speaker authorization is misconfigured, another person could be treated as allowed to trigger voice commands.
The skill explicitly uses speaker identity to grant command permissions. This is purpose-aligned, but mistakes in the owner or approved speaker registry could allow the wrong person to issue voice commands.
Owner (`is_owner: true`): Full command access, always authorized; Approved (`approved: true`): Can trigger wake word commands
Review the speaker list after setup, approve only trusted people, and periodically verify owner and approved entries.
Stored speaker profiles may reveal who participated in conversations and can affect future voice-command authorization.
The skill maintains persistent speaker profile data that can influence later identity and authorization decisions. This is expected for a speaker-ID skill, but the stored data should be kept accurate and protected.
Located at `percept/data/speakers.json` ... Speaker profiles track first/last seen timestamps and authorization status
Protect access to the speaker registry and dashboard, remove outdated profiles, and correct mislabeled speakers promptly.