Agent Blueprint
Coding · tested 2026-08-16 · re-test due 2026-11-16 · by the Hlido desk, not the vendor
In short: A cleanly-designed CLI + MCP server that delivers implementation 'blueprints' to coding agents as Agent Skills directories — but the whole thing is gated behind a commercial engagement, so almost nothing about it is independently verifiable.
6 PASS · 1 FAIL of 7 public-surface claims
Quick answer
Agent Blueprint scores 64/100 (FADING) on Hlido’s independent, hands-on test (reviewed 2026-08-16). FADING (64) because the agent-integration design is genuinely good — a proper CLI + built-in MCP server, adherence to the Agent Skills open standard, and a clean, well-named tool surface — but the entire product is gated Pricing was not findable on the public surface when tested.
Agent Blueprint's developer surface does the agent-integration mechanics correctly. It ships an npm package ('npm install -g agentblueprint') that is both a CLI and a built-in MCP server, and it commits to the Agent Skills open standard — every blueprint exports as a SKILL.md-anchored directory that Claude Code, Cursor, Codex and other Agent-Skills-aware agents auto-discover. The MCP server exposes a sensible, well-named tool set (list_blueprints, get_blueprint, download_blueprint, get_use_case, get_business_case, get_implementation_plan, get_implementation_spec, get_business_profile) and the config snippet is a clean 'npx -y agentblueprint serve'. For an agent that needs structured implementation context for a platform build (the example given is ServiceNow), this is exactly the right shape of integration. The problem for an independent rating is that the product is not open: the page states plainly that 'API tokens are issued with an engagement; if your organization works with us, you're ready to go.' In other words, the npm package is a thin client for a consulting/platform engagement — without an engagement there is no token, no blueprints, and nothing to actually run. The captured surface is correspondingly sparse: no pricing, no self-serve access, no sample blueprint to inspect, no evidence of the blueprints' quality or coverage, and no track record visible. So while the plumbing looks professional and genuinely agent-native, Hlido can only vouch for the integration design, not for the value of what flows through it. This is a review of a gate, not of the product behind it.
Why FADING
FADING (64) because the agent-integration design is genuinely good — a proper CLI + built-in MCP server, adherence to the Agent Skills open standard, and a clean, well-named tool surface — but the entire product is gated behind a commercial engagement (API tokens issued only 'with an engagement'), the public surface is thin, and there is no self-serve access, pricing, sample blueprint, or independent evidence of the blueprints' quality. Confidence is low: most of what would determine the product's actual worth sits behind the gate and cannot be verified from the surface. Not STEADY because so little is inspectable; not lower because what IS shown (the integration mechanics) is competent and honestly described.
Public-surface checklist
- PASS Homepage loads (required)
- PASS Primary value prop (required) — 'Bring Blueprints Into Your Workflow ... Every blueprint exports as an Agent Skills directory, an open standard for giving coding agents structured context'
- PASS Cta present (required) — 'npm install -g agentblueprint' + 'Talk to Us' / 'Contact'
- PASS Docs present (required) — CLI usage, MCP server config, and export-from-platform steps documented on the developers page
- PASS Integrations documented (required) — MCP server config for Claude Code, Cursor, Codex and other MCP-compatible agents; Agent Skills auto-discovery
- FAIL Pricing or access — No pricing and no self-serve access — API tokens are 'issued with an engagement'
- PASS Auth data handling — API-token auth on first CLI use; access bound to a commercial engagement
What we saw
2 screenshots captured by the Hlido engine during the reviewed run (run-21520bafe2496f7c-agentblueprint-ai). Our own captures — not vendor marketing material.
What it does well
- Agent-native integration done right: an npm-installed CLI plus a built-in MCP server ('npx -y agentblueprint serve') for Claude Code, Cursor, Codex and other MCP-compatible agents
- Commits to the Agent Skills open standard — blueprints export as SKILL.md-anchored directories that Agent-Skills-aware coding agents auto-discover
- Clean, sensibly-named MCP tool set (list/get/download blueprint, get use-case / business-case / implementation-plan / implementation-spec / business-profile)
- Honest about the access model — it states up front that API tokens come with an engagement, rather than implying self-serve availability
What it fails at
- Gated behind a commercial engagement: API tokens are 'issued with an engagement', so without becoming a client there is no token, no blueprints and nothing to run
- Thin public surface — no pricing, no self-serve signup, no sample blueprint to inspect, and no evidence of blueprint quality or coverage
- Nothing about the actual deliverable (the implementation specs) is independently verifiable from the surface; Hlido can only assess the integration layer
- No visible track record, customer evidence, or adoption signal on the captured page
Red flags
- Core functionality is inaccessible without a commercial engagement (API tokens 'issued with an engagement'), and the public surface offers no pricing, sample, or self-serve path — so the product cannot be independently evaluated before purchase.
Best for
- Organizations already in (or entering) a delivery engagement with Agent Blueprint who want their coding agents to pull structured implementation specs in-editor
- Teams standardizing on the Agent Skills format who value a first-class MCP + CLI path to structured platform-implementation context
- Platform-implementation work (e.g. ServiceNow) where a vetted implementation spec delivered as an Agent Skills directory saves setup time
Not recommended for
- Anyone wanting a self-serve tool they can try today — access requires a commercial engagement and an issued API token
- Buyers who need to inspect pricing, sample output, or independent quality evidence before committing
- Individual developers or small teams without an engagement relationship — there is no free or open path to the blueprints
- Use cases needing an open, inspectable knowledge base rather than a gated, engagement-bound one
Pricing & access
- Pricing findable on the public surfaceFAIL No pricing and no self-serve access — API tokens are 'issued with an engagement' (tested 2026-08-16)
Derived from Hlido-held evidence only (engine checklist + editorial text); quotes are verbatim from the scorecard; not vendor-supplied; re-derived daily. Verify current prices on the vendor's pricing page. Last verified 2026-08-16.
Related agents
Agent relevance
API CLI MCP
Agentic-Commerce Readiness 59/100 · INTEGRABLE
Independent readiness for agent delegation & transaction. How it’s scored · check live
A coding agent adds the MCP server via 'npx -y agentblueprint serve' (Claude Code / Cursor / Codex / MCP-compatible) and gains tools to list, fetch and download blueprints as Agent Skills directories; the CLI ('agentblueprint list' / 'download <id> --platform ...') does the same from the terminal. Both paths require an API token, which is issued only with a commercial engagement — so the integration is agent-native but not usable without becoming a client.
Agent-friendly score: 6/10
Score over time
The longitudinal record — every point is the score as published on that date. Raw series.
Evidence
- npm package 'agentblueprint' includes both a CLI and a built-in MCP server; blueprints export as Agent Skills directories auto-discovered via SKILL.md — source (2026-08-16) verified
- MCP server config is 'npx -y agentblueprint serve' and exposes tools list_blueprints, get_blueprint, download_blueprint, get_use_case, get_business_case, get_implementation_plan, get_implementation_spec, get_business_profile — source (2026-08-16) verified
- API tokens are issued with an engagement ('if your organization works with us, you're ready to go') — access is gated, not self-serve — source (2026-08-16) verified
- Quality, coverage and value of the actual blueprints / implementation specs — source (2026-08-16)

