surrealfs

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.

ConcernHigh Confidence
ASI05: Unexpected Code Execution
What this means

If exposed to untrusted prompts or inputs, the agent could be induced to run host commands through the pipe feature.

Why it was flagged

The skill documents a virtual-filesystem feature that can execute host commands. In an AI-agent context, that is broader than isolated virtual file operations and needs explicit containment.

Skill content
**Pipe commands**: The Rust core supports `curl URL > /path` syntax for content ingress. This executes the pipe source command on the host.
Recommendation

Do not expose pipe commands to autonomous or untrusted agents by default; require explicit user approval, allowlist commands/URLs, and run the tool in a container or sandbox.

What this means

The installed packages could contain behavior not visible in this review.

Why it was flagged

The runnable implementation comes from external package registries, while the supplied artifact set contains only SKILL.md and no reviewed code or lockfile.

Skill content
cargo install surrealfs ... pip install surrealfs-ai
Recommendation

Verify the package owners, source repository, versions, and hashes before installing; pin known-good versions where possible.

What this means

Over-privileged database credentials could let the agent read or modify more SurrealDB data than intended.

Why it was flagged

Remote SurrealDB access uses credentials. This is purpose-aligned and the skill advises least privilege, but it is sensitive authority and is not reflected in the registry-level credential declarations.

Skill content
SURREAL_USER ... sensitive: true ... SURREAL_PASS ... sensitive: true
Recommendation

Use dedicated least-privilege SurrealDB users scoped to one namespace/database, and avoid root credentials outside a local test environment.

What this means

Stored files or instructions may influence later agent sessions or be visible to other agents sharing the database.

Why it was flagged

The skill is designed to retain and reuse content across sessions and potentially across agents, which is useful but can also preserve sensitive or untrusted data.

Skill content
persistent, queryable virtual filesystem backed by SurrealDB ... content search across sessions ... Multi-agent shared file access
Recommendation

Separate projects into distinct namespaces/databases, periodically review stored content, and avoid storing secrets or untrusted instructions unless you have cleanup and access controls.

What this means

If telemetry is enabled or the HTTP service is exposed beyond localhost, operational data or agent tools may be reachable outside the intended local environment.

Why it was flagged

The Python agent introduces HTTP and telemetry/provider data flows. The defaults and opt-out guidance are disclosed, but users must preserve those boundaries.

Skill content
Telemetry via Pydantic Logfire (OpenTelemetry) ... HTTP hosting (default port 7932, bound to 127.0.0.1)
Recommendation

Keep the HTTP server bound to localhost unless protected by authentication and TLS, and disable or audit telemetry before processing sensitive data.