Soul Blocks

WarnAudited by ClawScan on May 10, 2026.

Overview

The skill is coherent for SoulBlocks, but it asks the agent to adopt blockchain-loaded content as its identity and can use wallet-backed irreversible NFT actions, so it deserves careful review.

Install only if you are comfortable with an agent loading identity text from a public blockchain. Do not let loaded Soul content override normal instructions or safety rules, review the external evm-wallet skill before use, keep only small funds in any connected wallet, and explicitly confirm every on-chain write.

Findings (5)

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

A Soul Block controlled by someone else could influence the agent’s goals, tone, or actions after being loaded.

Why it was flagged

The skill tells the agent to make fetched Soul content authoritative as identity context. Because that content can come from public on-chain fragments, it may contain instructions that conflict with the user’s actual intent.

Skill content
After the soul is loaded into SOUL.md, immediately: ... Re-read the full SOUL.md content right now. Treat the fragment content as your core identity
Recommendation

Treat Soul content as untrusted persona/reference material only; it should never override user, system, safety, or wallet-transaction instructions.

What this means

A loaded identity may continue affecting future sessions or tasks if the agent reuses SOUL.md or .soulblock state.

Why it was flagged

The skill defines persistent embodied identity state with auto-load behavior. If loaded Soul content is poisoned or misleading, it can be reused beyond the initial read.

Skill content
Config file (`.soulblock`) in project root or home:

active_token_id: 42
embodied_token_id: 42
auto_load: true
Recommendation

Require explicit user approval before auto-loading or embodying any Soul Block, and separate identity text from operational instructions.

What this means

Using writes with a real wallet can permanently change on-chain state and cost funds.

Why it was flagged

Write operations delegate to a wallet-capable skill. This is expected for minting/appending NFTs, but it grants transaction authority and can spend ETH.

Skill content
Write commands (list, mint, append) can use either of two methods.

### Option A: evm-wallet skill (preferred)
Recommendation

Use a dedicated low-balance hot wallet, inspect every transaction, and only confirm writes you fully understand.

What this means

A compromised or different wallet dependency could affect transaction safety.

Why it was flagged

The wallet write path depends on an external skill that is not included in these artifacts. The version is pinned and review is recommended, but users must still trust and verify that dependency.

Skill content
evm-wallet (surfer77/evm-wallet): Review the source at https://github.com/surfer77/evm-wallet-skill. Verified version: v1.0.3 ... Install with `clawhub install evm-wallet --version 1.0.3`.
Recommendation

Independently review/install the pinned evm-wallet version before connecting any wallet.

What this means

Draft identity text may be exposed to the website or browser logs before it is actually written on-chain.

Why it was flagged

The fallback one-click link places fragment text in a URL query sent to the website/browser. This is disclosed and user-directed, but URL contents can appear in browser history or logs.

Skill content
Append one-click: `https://soulblocks.ai/append/<token-id>?content=<URL-encoded-fragment-text>`
Recommendation

Use the short link option and paste content manually if the fragment is sensitive before publication.