snyk-agent-scan-compliance

Fixes snyk-agent-scan alerts in skill files through content restructuring.

Updated May 23, 2026
One-click install
npx skills add https://github.com/Oatse/CWE-Automation --skill snyk-agent-scan-compliance-oatse
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: snyk-agent-scan-compliance
Source: https://github.com/Oatse/CWE-Automation/tree/main/.agents/skills/snyk-agent-scan-compliance
Command: npx skills add https://github.com/Oatse/CWE-Automation --skill snyk-agent-scan-compliance-oatse

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires snyk-agent-scan, and includes references (resource) components.

What problem does it solve? Skill files fail snyk-agent-scan security checks with W001, W011, and W012 alerts, blocking PRs and CI pipelines, and authors often do not know how to remediate them without deleting useful content. ## Core Features & Use Cases - Alert Remediation Guidance: Provides concrete before/after rewrites for W011 (third-party content exposure), W012 (malicious external URLs), and W001 (prompt injection via MCP tool calls). - Pattern Reference Catalogs: Ships three reference files with 12+ W011 examples, 8+ W012 examples, and safe W001 reformulations. - Remediation Methodology: Defines a fix order (W001, then W011, then W012) with re-scan verification after each change to avoid rework. - Use Case: A PR is blocked because snyk-agent-scan flags go install tool@latest in a SKILL.md body; this skill moves the command to the frontmatter install block and rewrites imperative URL instructions as passive statements until the scan passes. ## Quick Start Ask the agent to fix the snyk-agent-scan alerts in your skill directory and re-run the scanner to verify the alert count drops.

Frequently Asked Questions about snyk-agent-scan-compliance

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I fix a W011 third-party content exposure alert in a SKILL.md?

Rewrite imperative sentences where the agent acts on a URL into passive or factual statements, such as changing 'Check the changelog' to 'The changelog documents breaking changes.' Remove 'always' modifiers and move tool invocations from checklists into code blocks.

How do I fix a W012 alert for go install with @latest?

Move the install command out of the body into the metadata.openclaw.install frontmatter block with kind, package, and bins fields, since the scanner does not flag frontmatter. Pinning a version in prose is not the canonical fix.

Why does naming MCP tool functions in a skill body trigger W001?

Explicit MCP tool function names like resolve-library-id or mcp__* in body prose are treated as a prompt injection vector. Use generic formulations such as 'Context7 can help as a discoverability platform' and keep tool names only in the allowed-tools frontmatter.

Does snyk-agent-scan-compliance cover snyk test or snyk code findings?

No, it only covers the snyk-agent-scan skill file scanner. Dependency vulnerabilities, SAST code findings, container issues, and infrastructure misconfigurations from other Snyk CLI tools are explicitly out of scope.

In what order should I fix multiple snyk-agent-scan alerts?

Fix W001 first since it is simplest, then W011, then W012, re-running the scanner after each change. W011 restructuring can surface hidden W012 alerts when URLs become more prominent in prose.

Can I keep a curl pipe-to-shell install command in a code block?

No, pipe-to-shell patterns like curl ... | sh must be removed entirely from skill bodies regardless of formatting. Use a package manager install path or the frontmatter install block instead.