Percept Summarize
ReviewAudited by ClawScan on May 10, 2026.
Overview
This skill is coherent for meeting summaries, but it automatically processes conversations and keeps searchable conversation/person data, including speaker profiles that never expire, with limited control details.
This skill appears designed for meeting-memory workflows, but install it only if you want conversations automatically summarized and stored. Confirm the listener consent model, local database location, dashboard access controls, retention/deletion process, and whether speaker profiles can be removed or set to expire.
Findings (3)
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.
Private or sensitive conversations could be summarized automatically once the listening component is running.
The skill processes and submits conversation transcripts automatically after a silence timeout, but the artifact does not describe user approval, opt-out, or limits on which conversations are processed.
Conversation ends (60s silence timeout) ... Sends transcript to OpenClaw for AI summarization
Use only if you are comfortable with automatic summarization, and look for controls that let you choose which conversations are captured, summarized, or skipped.
Conversation content and people/relationship data may remain available to the agent or dashboard long after the meeting, increasing privacy and context-reuse risk.
The skill creates persistent searchable memory from conversations, including entity relationships and non-expiring speaker profiles, without describing deletion, review, exclusion, or reuse boundaries.
Stores summary + entities in SQLite ... Entities linked via relationship graph ... Speaker profiles: never expire
Before installing, confirm where the SQLite/LanceDB data is stored, how to delete it, whether speaker profiles can expire, and how the agent is allowed to reuse stored conversation context.
The safety of this skill also depends on the separate listener and OpenClaw CLI setup.
This instruction-only skill depends on other runtime components that are not included in the reviewed artifacts, so their behavior and permissions cannot be assessed here.
Requirements: percept-listen skill installed and running; OpenClaw agent accessible via CLI
Review the percept-listen skill and OpenClaw CLI configuration before relying on this summarization skill.