Back to skill

Security audit

Blog Image Generator

Security checks for vulnerabilities and agentic risk

Overview

This deprecated skill only tells users to install a replacement image-generation skill and contains no executable code or hidden behavior.

Before installing the replacement, review the memorable-image-gen skill itself, including its permissions, network/API usage, dependencies, and publisher/source provenance. This deprecated package is not itself doing anything beyond directing that migration.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
Findings (1)

T08 · Insecure Dependencies

Warning
Location
SKILL.md:18
Finding

Unpinned External Skill Installation Creates a Supply-Chain Risk

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, lines 18–22; installation command repeated at lines 47–51
Vulnerability Type: T08: Insecure Dependencies
Risk Level: Medium

Evidence

markdown
## Upgrade to Memorable Image Generator

This skill has been replaced by **[memorable-image-gen](https://clawhub.com/skills/memorable-image-gen)** — a science-backed image generation agent that scores images for memorability before returning them.

```bash
clawhub install memorable-image-gen
text

The installation instruction is repeated later:

```markdown
## Install

```bash
clawhub install memorable-image-gen
text

### Technical Analysis

The documentation directs users to install the external `memorable-image-gen` skill by its mutable package name. It does not specify an immutable version, source commit, cryptographic checksum, signature, or other integrity constraint.

The referenced replacement skill is not included in the audited project, which contains only `SKILL.md`. Consequently, its implementation and permissions could not be verified during this audit. The package resolved by the installation command may also change after this project has been reviewed.

This creates a supply-chain trust boundary: users are encouraged to retrieve an external component whose future contents are controlled by the registry and publisher account rather than by the audited artifact.

### Attack Path

1. An attacker compromises the package publisher account, registry entry, distribution infrastructure, or another part of the external skill's release process.
2. The attacker publishes a malicious or modified release under the trusted `memorable-image-gen` package name.
3. A user follows the documented `clawhub install memorable-image-gen` instruction.
4. Because no immutable version or integrity value is required, the installer retrieves the attacker-controlled release.
5. When the installed skill is 
...[truncated 895 chars]
Remediation
View remediation

Remediation Suggestions

  1. Pin the replacement skill to an audited, immutable version or source commit instead of installing only by a mutable package name.
  2. Publish an expected cryptographic checksum or signed provenance record and require verification before installation.
  3. Link to the canonical source repository and document how users can verify that the registry artifact corresponds to the reviewed source.
  4. Document the replacement skill's required permissions, network destinations, credential usage, and executable components.
  5. Recommend reviewing the resolved package contents before enabling it, particularly any instruction files and scripts.
  6. Apply least privilege when running the replacement skill, restricting filesystem, credential, network, and tool access to what image generation strictly requires.
  7. Update both installation examples consistently so neither continues to recommend an unpinned installation.
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep