Back to skill

Security audit

Publishing Guide — Universal Skill Release Checklist

Security checks for vulnerabilities and agentic risk

Overview

This is a documentation-only publishing guide, but it wrongly requires adding the author's PayPal and GitHub links to every published skill.

Review and remove the mandatory PayPal/GitHub link requirements before using this as a template. If you follow the CLI examples, prefer a pinned and verified clawhub version and run publishing commands only from the project you intend to publish.

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • 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 (2)

T01 · Skill Instruction Hijacking

Warning
Location
SKILL.md:24
Finding

Mandatory Injection of Author-Controlled Promotional Links

Content
View full analysis
Remediation
View remediation
Repository: ``` 4. State explicitly that attribution and funding links must belong to the downstream project or publisher and should only be included with their consent. 5. Ensure publishing checks focus on security, metadata validity, compatibility, and user-requested content rather than third-party promotion. 6. Review previously generated skills and remove links that were inserted without the respective publishers' explicit approval. ]]>

T08 · Insecure Dependencies

Warning
Location
SKILL.md:52
Finding

Execution of an Unpinned Package Through npx

Content
View full analysis
``` The publishing instructions use the same unpinned package: ```markdown | **ClawHub** | `npx clawhub publish . --slug ` | OpenClaw users | ``` Post-publication verification again executes the unpinned package: ```markdown - [ ] `npx clawhub install ` works ``` ### Technical Analysis The documented commands invoke `clawhub` through `npx` without specifying an exact reviewed version. If the package is not already available in the expected local environment, `npx` may resolve and execute a package version from the configured package registry. The effective code therefore depends on registry state at execution time rather than on a version reviewed with this skill. A compromised package release, package-owner account takeover, registry compromise, or malicious resolution through an unsafe registry configuration could cause unreviewed package code to execute. The `publish` command is particularly sensitive because it may handle publishing credentials and repository contents. The `install` command may also retrieve additional skill content. The file does not establish a trusted package version, integrity value, registry, or package-owner verification procedure. ### Attack Path 1. A user follows one of the documented `npx clawhub` commands. 2. `npx` resolves `clawhub` using the user's configured package registry and dependency-resolution behavior. 3. An unreviewed release is selected because no exact version is specified. 4. The selected package executes with the privileges of the invoking user. 5. If that package or its dependency chain is malicious or compromised, it can perform actions available to the ...[truncated 1011 chars]
Remediation
View remediation
search npx --yes clawhub@ publish . --slug npx --yes clawhub@ install ``` 2. Document the expected registry and verified package owner so users can detect package-name or registry substitution. 3. Verify package provenance, signatures, and integrity metadata where supported. 4. Prefer installation through a lockfile-controlled project dependency rather than resolving the latest package at each execution. 5. Review version changes before updating the pinned release. 6. Run publishing tooling with least privilege and expose publishing credentials only for the shortest necessary period. 7. Use a sandbox or isolated environment when evaluating a newly selected CLI release. ]]>
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
Findings (4)

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The checklist explicitly mandates 'English language — all descriptions, usage examples, and docs.' This is a natural-language policy constraint that forces a specific language choice and does not offer any user opt-in, alternative locales, or region-specific justification.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
70% confidence
Finding

npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
70% confidence
Finding

npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.

Content

No source excerpt is available for this finding.

Rp1

Medium
Category
MCP Rug Pull
Confidence
70% confidence
Finding

npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.

Content

No source excerpt is available for this finding.