ClawGraph

Security checks across static analysis, malware telemetry, and agentic risk

Overview

ClawGraph is a coherent memory skill, but it persistently stores user facts and uses an OpenAI-compatible API, so it should be treated as privacy-sensitive.

Install only if you want automatic long-term memory. Before using it, confirm you are comfortable with durable facts being stored under ~/.clawgraph/data and processed by your configured OpenAI-compatible provider; avoid sharing sensitive personal or confidential business information unless you intend it to be remembered.

Static analysis

No static analysis findings were reported for this release.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Risk analysis

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

Facts about you, your projects, or your team can persist and be reused in later conversations, including if a statement later becomes outdated or was something you did not intend to remember long term.

Why it was flagged

The skill intentionally creates persistent cross-conversation memory from user statements.

Skill content
Proactively store durable user facts without waiting for an explicit memory command ... Data stored at `~/.clawgraph/data` — survives restarts
Recommendation

Use this only if you want automatic durable memory; periodically inspect the graph and avoid sharing sensitive or confidential details unless you are comfortable with them being stored.

What this means

The configured provider account may be used for model calls, which can affect billing, usage logs, and provider-side data handling.

Why it was flagged

The skill requires a provider credential and can be pointed at an OpenAI-compatible endpoint, which is expected for its extraction workflow.

Skill content
`OPENAI_API_KEY` is required. `OPENAI_BASE_URL` is optional for other OpenAI-compatible endpoints.
Recommendation

Use a dedicated API key where possible, monitor usage, and set OPENAI_BASE_URL only to an endpoint you trust.

What this means

User facts may be sent to the configured OpenAI-compatible provider for extraction, not just stored locally.

Why it was flagged

Facts selected for storage are processed through an LLM provider interface.

Skill content
Each fact is automatically decomposed into entities and relationships using ... one LLM call ... OpenAI-compatible APIs today via the OpenAI SDK.
Recommendation

Review the provider's data handling policy and avoid storing facts that should not leave your environment.

What this means

Installing the skill will install and run code from the external package, which was not directly visible in the supplied files.

Why it was flagged

The executable comes from an external pinned package; the provided artifact set did not include the package code for static review.

Skill content
uv | package: clawgraph==0.1.3 | creates binaries: clawgraph
Recommendation

Install from a trusted package source, verify the package/version and homepage, and consider using an isolated environment.