# io.github.cdeust/ai-architect-mcp-spec (mcpb · ai-architect-mcp-spec.mcpb)

PRD verification and host-driven generation with multi-judge consensus across 17 MCP tools.

- Trust score: 50/100 (low)
- Change this week: +6
- Registry status: active
- Liveness: live
- Owner verified: no
- Last scored: 2026-08-22

## Components

- mcpb · `ai-architect-mcp-spec.mcpb`: 50/100 (this document), [markdown](https://verifymcp.io/servers/cdeust-ai-architect-mcp-spec/https-github-com-cdeust-ai-architect-mcp-spec-releases-download-v0-8-0-ai-archit.md), [page](https://verifymcp.io/servers/cdeust-ai-architect-mcp-spec/https-github-com-cdeust-ai-architect-mcp-spec-releases-download-v0-8-0-ai-archit)

## Channel facts

- Registry: `mcpb`
- Package: `https://github.com/cdeust/ai-architect-mcp-spec/releases/download/v0.8.0/ai-architect-mcp-spec.mcpb`
- Version: `0.8.0`
- Transport: `stdio`

## Trust breakdown

How this component scores in each security and reliability category. Every signal is checked automatically from public evidence about the published package, including repeated runs of it in an isolated sandbox, and we only credit what we can confirm. Scores are 0–100 per category. Scoring method: https://verifymcp.io/docs/scoring (what has changed: https://verifymcp.io/docs/scoring/changelog)

Scored 2026-08-22.

- **Supply Chain Security**: 13/100
  - Malware scan not yet available for this package.
  - Known CVEs could not be checked: this artifact ships no SBOM, so there is no dependency list to read. Publishing one would let us assess it.
  - No install/post-install scripts declared.
  - Dependency health could not be checked: this artifact ships no SBOM, so there is no dependency list to read. Publishing one would let us assess it.
- **Provenance & Transparency**: 48/100
  - Source repository is publicly reachable at the declared URL.
  - Provenance check failed: no build-provenance attestation is published.
  - Clear OSI-approved license (MIT).
  - Actively maintained (last published 11 days ago).
  - Publishes a security disclosure policy (SECURITY.md).
- **Schema Quality & AI Usability**: 83/100
  - 100% of prompts and resources have a non-trivial description (not blank, and not just the item's name).
  - AI-judged instruction clarity (good).
  - Tool/resource definitions use about 1453 tokens (~85/item across 17 items; 17 tools + 0 resources), lean.
  - Usage-examples check failed: none of the tools include examples.
- **Stability & Change Management**: 40/100
  - Stability observed for 12 of 30 days with no destabilising changes; credit accrues until the full window elapses.
- **Tool Coverage**: 84/100
  - 100% of tools have a non-trivial description (not blank, and not just the tool's name).
  - 52% of tool parameters carry a description.
- **Capabilities**: 100/100
  - Implements a supported MCP spec version (2025-11-25); the latest is 2026-07-28.

## Install

- Download bundle: `https://github.com/cdeust/ai-architect-mcp-spec/releases/download/v0.8.0/ai-architect-mcp-spec.mcpb`

## Changelog

Every change recorded for this component, newest first. Days that predate change tracking, or that we cannot explain, say so: "we were watching and nothing happened" and "we were not watching" are different claims.

### 2026-08-22 (score 50, +1)

No change was recorded against any check on this day. Stability & Change Management went from 37 to 40. That category is still filling its 30-day observation window: 11 days of observed history at the previous scan, 12 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-08-21 (score 49, +1)

No change was recorded against any check on this day. Stability & Change Management went from 33 to 37. That category is still filling its 30-day observation window: 10 days of observed history at the previous scan, 11 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-08-18 (score 48, +4)

- [functional improvement] Stability: unverified → 0.27

### 2026-08-11 (score 44, +4)

- [functional] We updated how we score, so this day's move reflects our rubric, not a change to the server

### 2026-08-10 (score 40)

First indexed and scored.

## MCP tools (17)

### `get_config` (~14 tokens)

Get the full skill configuration

### `read_skill_config` (~21 tokens)

Read the SKILL.md content that drives PRD generation

### `check_health` (~18 tokens)

Check system health — verify all components are accessible

### `get_prd_context_info` (~33 tokens)

Get configuration for a specific PRD context type

Input parameters:

- `context` (string, required): The PRD context type

### `list_available_strategies` (~19 tokens)

List thinking strategies available to the pipeline.

### `validate_prd_section` (~68 tokens)

Run deterministic Hard Output Rules validation on a single PRD section. Returns violations found — zero LLM calls, pure regex/parsing.

Input parameters:

- `content` (string, required): The markdown content of the PRD section
- `section_type` (string, required): The type of PRD section being validated

### `validate_prd_document` (~53 tokens)

Run full document validation including cross-section checks (SP arithmetic, AC numbering, FR-AC coverage, test traceability). Returns comprehensive validation report.

Input parameters:

- `sections` (array, required): Array of PRD sections to validate

### `get_quality_history` (~33 tokens)

Get historical PRD quality scores from the evidence repository

Input parameters:

- `limit` (integer): Maximum number of records to return

### `get_strategy_effectiveness` (~40 tokens)

Get strategy performance data — actual vs expected improvement, compliance rate

Input parameters:

- `min_executions` (integer): Minimum executions required to include a strategy

### `coordinate_context_budget` (~84 tokens)

Calculate token budget allocation for PRD generation. Returns per-section retrieval limits for Cortex recall, generation budgets, and section-specific query templates. Call this BEFORE starting section generation.

Input parameters:

- `completed_sections` (array): Section types already generated
- `context_window_size` (integer): Total context window size in tokens
- `prd_context` (string, required): The PRD context type

### `map_failure_to_retrieval` (~58 tokens)

When validate_prd_section returns violations, call this to get corrective Cortex recall queries. Closes the validation→retrieval feedback loop so retries use better context.

Input parameters:

- `violations` (array, required): Violations from validate_prd_section

### `start_pipeline` (~124 tokens)

Initialize a new PRD pipeline run. Returns run_id and the first NextAction the host must execute.

Input parameters:

- `codebase_path` (string): Absolute path to the codebase. Triggers index_codebase via ai-architect-mcp-codebase.
- `feature_description` (string, required): What the PRD is about — passed to all prompts
- `skip_preflight` (boolean): If true, skip the preflight step that probes Cortex (and ai-architect when codebase_path is set). Default false. Use only when you accept degraded section generation without persistent memory recall.

### `submit_action_result` (~37 tokens)

Feed an ActionResult to the pipeline runner; receive the next NextAction.

Input parameters:

- `result` (required)
- `run_id` (string, required)

### `get_pipeline_state` (~189 tokens)

Read the current pipeline state by run_id. format:'summary' (default) returns the lightweight envelope; format:'full' returns the whole state, bounded to the Claude Code 100,000-char MCP response budget by shedding least-relevant detail first (observable __bounded markers; full grounding re-fetchable via format:'grounding'); format:'grounding' returns the codebase_grounding (+ prd_validation when it fits) blobs format:'full' sheds first; format:'validation' returns prd_validation alone (the blob format:'grounding' sheds when the pair overshoots); format:'action' returns the UNBOUNDED last action emitted for this run (including full spawn_subagents prompts) — the recovery path when a start_pipeline/submit_action_result response carries a __bounded marker on its action.

Input parameters:

- `format` (string)
- `run_id` (string, required)

### `plan_section_verification` (~69 tokens)

Extract claims from a PRD section and select judges. Returns JudgeRequest[] the host must execute via Agent tool in parallel.

Input parameters:

- `codebase_excerpts` (array)
- `content` (string, required)
- `memory_excerpts` (array)
- `section_type` (string, required)

### `plan_document_verification` (~50 tokens)

Same as plan_section_verification but across all sections of a document.

Input parameters:

- `codebase_excerpts` (array)
- `memory_excerpts` (array)
- `sections` (array, required)

### `conclude_verification` (~363 tokens)

Aggregate JudgeVerdict[] from spawned subagents into a VerificationReport (consensus + dissent). IMPORTANT: omitting claim_types when a reliability repository is open suppresses observation flushing for this batch — the calibration data will be missing for these runs (one-sided censoring).

Input parameters:

- `claim_types` (object): Map of claim_id → claim_type. When provided, enables per-(judge × claim_type) reliability lookup. Omit to fall back to per-agent scalar priors. Source: derive from plan_section_verification / plan_do…
- `claims` (array): OPTIONAL. Pass the Claim objects from the corresponding plan_section_verification / plan_document_verification response if you want oracle-based ground truth (breaks Curie A2 annotator-circularity fo…
- `consensus_strategy` (string)
- `run_id` (string): Pipeline run_id — required for calibrated Bayesian reliability weights (CC-3 control-arm seam uses this to partition treatment vs control runs). When absent, falls back to Beta(7,3) prior for all jud…
- `scope` (string)
- `section_type` (string)
- `verdicts` (array, required)

## Diagnostics

Captured diagnostic sections: Provenance. The full working is on the page: https://verifymcp.io/servers/cdeust-ai-architect-mcp-spec/https-github-com-cdeust-ai-architect-mcp-spec-releases-download-v0-8-0-ai-archit#diagnostics

## Score history

- 2026-08-22: 50
- 2026-08-21: 49
- 2026-08-20: 48
- 2026-08-19: 48
- 2026-08-18: 48
- 2026-08-17: 44
- 2026-08-16: 44
- 2026-08-15: 44
- 2026-08-14: 44
- 2026-08-13: 44
- 2026-08-12: 44
- 2026-08-11: 44
- 2026-08-10: 40

## Links

- Repository: https://github.com/cdeust/ai-architect-mcp-spec
- Website: https://ai-architect.tools/
- Changelog RSS feed: https://verifymcp.io/servers/cdeust-ai-architect-mcp-spec/https-github-com-cdeust-ai-architect-mcp-spec-releases-download-v0-8-0-ai-archit.xml
- Changelog JSON feed: https://verifymcp.io/servers/cdeust-ai-architect-mcp-spec/https-github-com-cdeust-ai-architect-mcp-spec-releases-download-v0-8-0-ai-archit.json
- HTML version of this page: https://verifymcp.io/servers/cdeust-ai-architect-mcp-spec/https-github-com-cdeust-ai-architect-mcp-spec-releases-download-v0-8-0-ai-archit