Percept Ambient

Continuously captures and summarizes ambient conversations to build a local knowledge graph for context-aware assistance without explicit commands.

This skill enables always-on passive capture of conversations and creates searchable transcripts and vectors — a high privacy and attack-surface item. Before installing, do the following: 1) Review and audit the referenced components (percept-listen, percept-summarize, LanceDB, any embedding provider) — inspect their source code and network behavior. 2) Confirm how transcripts are stored: where files live, whether they are encrypted at rest, and whether TTL/auto-purge is actually implemented and enforced. 3) Verify whether NVIDIA NIM embeddings require remote API access, credentials, or GPU hardware; the skill does not declare these needs. 4) Ensure the dashboard is bound to localhost, protected by authentication, and not inadvertently exposed to the LAN/Internet. 5) Require explicit, revocable user consent and per-session controls for listening; log and surface when recording/transcription occurs. 6) If you cannot audit the complementary skills and runtime components, avoid installing this — consider alternatives that explicitly document consent, storage, encryption, and data-flow guarantees.

Type: OpenClaw Skill Name: percept-ambient Version: 1.0.0 The skill `percept-ambient` is classified as suspicious due to its core functionality involving continuous ambient speech capture and transcription, which inherently handles highly sensitive personal data. While the `SKILL.md` explicitly states that data is stored locally and no audio is retained (only transcripts), the continuous processing of private conversations represents a significant privacy risk. Additionally, the skill exposes a local API and dashboard on `http://localhost:8960`, which, while intended for monitoring, introduces a potential attack surface for local privilege escalation or other vulnerabilities if not robustly secured. There is no evidence of intentional data exfiltration, malicious execution, or prompt injection attempts against the agent within the provided files, but the high-risk capabilities warrant a 'suspicious' classification.