User-hosted MCP server that brings infracodebase's compliance, rulesets, and governance to your local agent.
Strong foundation with 16 well-named tools following verb_noun convention (list_, get_, create_, update_, trigger_). All tools have descriptions (avg 180 chars, within 10-1024 baseline). Input schemas present and typed for all tools. However, output schemas are not documented in source, LLMs cannot see what fields to expect from responses. Parameter descriptions are present but inconsistent in depth; some lack format/constraint guidance. Error handling is minimal, no recovery guidance or actionable error messages visible. Tool descriptions are functional but could be more LLM-optimized (e.g., 'When to use this' and 'What it returns' sections missing). No tool annotations (readOnlyHint/destructiveHint) despite clear risk classifications in metadata.
Create a workspace, optionally linking it to a repo (any provider). The repo link is only sent when all three repo fields are present.
The system prompt and conventions the CI compliance evaluator uses.
Summary of a compliance evaluation; latest when no ref is given. `branch` scopes "latest" to that branch (server falls back to the workspace's default branch, then to the most recent completed evaluation on any branch); ignored when `ref` is given.
Get detailed information about a specific ruleset.
Full workspace context, resolved from either an explicit workspace_id or a repo_url in a single server-side request. Response `status` is one of `linked`, `unlinked` (no workspace matches), `no_access` (a workspace exists but the caller can't see it), or `ambiguous` (the repo matches workspaces in more than one enterprise the caller can see).
Output schemas not documented. LLMs cannot see what fields responses contain, forcing them to guess at downstream tool parameters and breaking tool chaining.
No tool annotations (readOnlyHint/destructiveHint/idempotentHint) despite clear risk classifications. Agents cannot distinguish safe reads from destructive writes without reading descriptions.
Error handling not visible in source. No recovery guidance, actionable error messages, or error classification (retryable vs user-fixable vs fatal).
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | C | 69 | 2026-07-28+ | v2 |
Link a workspace to a repo (any provider) for compliance evaluations on push.
Per-rule findings from a compliance evaluation.
Rulesets, MCP servers, and workflows available in an enterprise.
List enterprises the caller belongs to.
The enterprise's approved reusable infrastructure modules.
Version-control connections (GitHub, GitLab, …) configured for an enterprise.
Repositories a version-control connection can reach.
List every ruleset relevant to a workspace — enterprise rulesets (required and optional, including optional ones the workspace hasn't opted into), workspace-owned, and the caller's personal rulesets.
List workspaces in an enterprise.
Trigger a compliance evaluation. With no scoping params, runs a full evaluation. Pass `ruleset_id`, `rule_id`, or `rule_ids` to scope the run to just those rules.
Update workspace resources (rulesets, MCP servers, workflows).
Parameter descriptions lack format/constraint guidance. E.g., 'ref' in get_compliance_evaluation says 'commit SHA or evaluation ID' but no regex/length constraints. 'status' in list_compliance_findings has no enum values listed.
Pagination not visible in list tools. list_enterprises, list_workspaces, list_workspace_rulesets, list_compliance_findings, list_enterprise_resources, list_modules, list_vcs_repos lack limit/offset/cursor parameters and total count in responses.