Single tool 'secure_coding_guidelines' has critical definition gaps. Input schema is empty object with no properties, violates pattern:constrained-input and pattern:tool-description. Tool description (67 chars) is generic and lacks actionable context for LLM selection. No parameter descriptions exist (schema has zero properties). No output schema documented. Tool is read-only but lacks idempotentHint annotation. No error handling guidance. Naming is acceptable (verb_noun form) but tool is essentially a static prompt delivery mechanism, not a functional agent tool.

• Add explicit input schema with typed properties. If tool accepts no parameters, document this clearly: 'inputSchema: {type: "object", properties: {}, required: []}' with description 'This tool requires no input parameters.'
• Expand tool description to 150 - 250 characters. Example: 'Returns OWASP-aligned secure coding guidelines for code generation. Call this before generating code to inject security best practices into the LLM context. Returns structured guidelines covering input validation, authentication, error handling, and cryptography.'
• Document output schema explicitly. Define return type as {type: 'object', properties: {guidelines: {type: 'string', description: 'OWASP-aligned secure coding guidelines...'}}}.
• Add idempotentHint: true to tool definition (via toolAnnotations in capabilities) to signal to LLM that repeated calls are safe.
• Consider redesigning tool to accept parameters (e.g., language, framework, threat_model) so LLM can request context-specific guidelines rather than a generic prompt. This increases utility and aligns with pattern:tool-composition.