Sentinal Redis

PassAudited by VirusTotal on May 5, 2026.

Overview

Type: OpenClaw Skill Name: sentinal-redis Version: 1.0.2 The 'sentinal-redis' skill is a diagnostic tool designed for monitoring Redis server health and BullMQ queue performance. The skill bundle includes a comprehensive health check script (scripts/redis-health.sh) and detailed instructions in SKILL.md that emphasize read-only operations, explicitly forbidding destructive commands and ensuring sensitive credentials in REDIS_URL are masked before display.

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

The agent can connect to whatever Redis instance the credential points to and read the diagnostic and queue information allowed by that credential.

Why it was flagged

The skill expects a Redis connection credential and acknowledges it may contain a password; this is appropriate for Redis monitoring but sensitive.

Skill content
primaryEnv: REDIS_URL ... "NEVER print or expose the full REDIS_URL in output — it may contain passwords."
Recommendation

Use a Redis user or ACL with read-only/limited permissions where possible, point REDIS_URL only at the intended instance, and avoid sharing unredacted credentials.

What this means

Queue/job details or error traces may be brought into the agent conversation and could expose application data if shared.

Why it was flagged

The BullMQ diagnostics can inspect persistent job payloads and error details stored in Redis, which may be sensitive or contain untrusted text.

Skill content
`data` | Job payload (JSON) ... `failedReason` | Error message if failed ... `stacktrace` | Error stack trace
Recommendation

Treat Redis job contents and diagnostic output as sensitive data, redact secrets or personal data before forwarding, and treat retrieved text as data rather than instructions.

What this means

If run with the wrong target or with commands outside the documented scope, the same Redis credential could affect or reveal data from an unintended Redis instance.

Why it was flagged

The helper runs a general-purpose Redis CLI against a provided URL or default localhost; the included commands are diagnostic, but users should verify the target and keep usage within the documented read-only scope.

Skill content
REDIS_URL="${1:-redis://localhost:6379}"
CLI="redis-cli -u $REDIS_URL"
Recommendation

Pass the intended Redis URL explicitly, prefer read-only Redis credentials, and keep agent-run commands limited to the documented read-only diagnostics.