Curated read-only Enji Guard surface for portfolio overview and repository audit reading. Authentication and operator controls belong to the CLI/runtime.
Two read-only tools with clear, concise descriptions (60 - 90 chars each). Both have input schemas with typed parameters and descriptions. However, output schemas are not documented in the visible code, parameter descriptions lack detail on format/constraints, and no error handling guidance is provided. Tool names follow verb_noun convention but are somewhat generic ('overview', 'audits'). Composition is sound, tools are single-responsibility and chainable via project/repo identifiers.
Read projects, repositories, scores, and project-level audit activity.
Read compact audit status and summaries; name audit selectors to include those Markdown report bodies.
Output schemas not documented. LLMs cannot infer what fields to expect from portfolio_overview or repo_audits responses, forcing them to guess at downstream field names and risk broken tool chains.
Parameter descriptions lack format/constraint details. 'project' is described as 'Exact project id or name filter (optional)' but does not specify format, length, or valid characters. 'sort' lacks enum values or allowed options. LLMs will hallucinate invalid values.
No error handling guidance. Tool descriptions do not explain what happens on invalid input, missing resources, or API failures. LLMs have no recovery path if a call fails.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 53 | 2026-07-28+ | v2 |
Tool names are generic. 'enji_portfolio_overview' and 'enji_repo_audits' do not clearly signal what data is returned or when to use each. More specific names like 'list_portfolio_projects' and 'get_repository_audit_status' would improve LLM disambiguation.
No pagination or result-limit documentation. If portfolio_overview or repo_audits return large lists, LLMs have no guidance on how to paginate or what the maximum result size is.