Agent teams, swarms, memory and hooks for Claude Code and Codex, served over MCP.
Prompt Optimizer is installed from its publisher's own source and answers where it runs, so this marketplace is not in the path of a single call. There is no address here to send one to, and a panel that pretended otherwise would be showing you an answer we made up. Install it and call it from your own client — the Installation tab has the entry for each one.
What it does
Optimize prompts and generate agent SOPs and skill packages with the Prompt Optimizer cloud API.
Quickstart
# 1 — add it to Claude Code
claude mcp add --scope user --env 'OPTIMIZER_API_KEY=${OPTIMIZER_API_KEY}' --transport stdio nivlewd1-prompt-optimizer-mcp '--' npx -y mcp-prompt-optimizer
# 2 — ask your agent for something these tools do
# optimize_prompt, get_quota_status, create_template and 21 more
It needs OPTIMIZER_API_KEY: this reads it from your environment, so export it first.
Prompt Optimizer costs nothing on mcprush: there is no plan to choose here, no cap to set and nothing mcprush can bill you for.
Collected from a public index. Nobody has claimed this account, so nothing here was written by its author — claim it if it is yours.
Where are you running it?
Every route below installs the same thing and ends at the same approval screen. This one runs on your machine: your client starts Prompt Optimizer as a process under your own user, with your files and your network, so the tool surface below is what it can do to your computer rather than to a server somewhere else. Its scan result is on the Security tab — read the surface before you approve it.
This is a public server: you run it yourself and this marketplace is not in the path. You need Node.js — the current LTS (24) is the safe choice; npx comes with it. Claude Code registers it in one command; --scope user makes it available in every project. Claude Code reads OPTIMIZER_API_KEY (Your Prompt Optimizer API key (starts with sk-opt-). A free key takes no card: https://promptoptimizer.xyz/dashboard) from your environment each time it starts the server, so the value stays out of its config: export it in your shell profile (macOS, Linux), or run setx OPTIMIZER_API_KEY … and open a new terminal (Windows). The '--' is in quotes so that PowerShell passes it on.
claude mcp add --scope user --env 'OPTIMIZER_API_KEY=${OPTIMIZER_API_KEY}' --transport stdio nivlewd1-prompt-optimizer-mcp '--' npx -y mcp-prompt-optimizerReconnect, or start a new session, and the tools appear in the model’s tool list.
It reaches a system of yours, so it needs your own credential rather than the publisher’s: OPTIMIZER_API_KEY — Your Prompt Optimizer API key (starts with sk-opt-). A free key takes no card: https://promptoptimizer.xyz/dashboard. You set it in your own client’s config, on your machine — this marketplace never holds it.
One config entry your client uses to start the process locally. A local server runs with your file system and your network, which is why it is priced without metering.
24 tools, with what each one reads, writes and reaches shown before you agree — the same list on every route above. Read the tool surface.
Tool surface
What the model actually sees. Descriptions are diffed on every release — see version history.
Read from its source code. A server that builds its tool list when it starts can list more tools than this.
🎯 Professional AI-powered prompt optimization with intelligent context detection, template auto-save, and comprehensive optimization insights
| Parameter | Type | Required | Notes |
|---|---|---|---|
| prompt | string | yes | The prompt text to optimize |
| goals | array | no | Optimization goals (e.g., 'clarity', 'conciseness', 'creativity', 'technical_accuracy', 'analytical_depth', 'creative_enhancement') |
| ai_context | string | no | The context for the AI's task (auto-detected if not specified with enhanced detection) |
| enable_bayesian | boolean | no | Deprecated: accepted for compatibility; has no effect on the hosted service |
| value_hierarchy | array | no | Ordered list of values/constraints the optimizer must respect. NON_NEGOTIABLE entries force LLM-tier routing and inject hard constraints into the system prompt. Example: [{label:'NON_NEGOTIABLE',descr |
| intent_frame | object | no | Question Method intent framing — steers optimization toward a specific angle, excludes off-topic territory, and defines what success looks like. Any non-null field floors routing to HYBRID tier minimu |
| reasoning_effort | string | no | How much reasoning to apply: 'minimal' biases toward faster/cheaper routing, 'standard' is the default, 'deep' biases toward the LLM tier for maximum analysis. Most useful when calling this tool progr |
| execution_shape | string | no | Execution style, independent of the tier the prompt would normally route to: 'direct' is single-pass, 'hybrid' adds rules+LLM verification, 'multi_agent' forces plan-and-execute with sub-agents regard |
📊 Check subscription status, quota usage, and account information with detailed insights
No parameter schema on file.
➕ Create a new optimization template.
| Parameter | Type | Required | Notes |
|---|---|---|---|
| title | string | yes | Title of the template |
| description | string | no | Description of the template |
| original_prompt | string | yes | The original prompt text |
| optimized_prompt | string | yes | The optimized prompt text |
| optimization_goals | array | no | Goals for this optimization (e.g., 'clarity', 'conciseness', 'creativity', 'technical_accuracy', 'analytical_depth', 'creative_enhancement') |
| confidence_score | number | yes | Confidence score of the optimization (0.0-1.0) |
| model_used | string | no | Model used for optimization |
| optimization_tier | string | no | Tier of optimization (e.g., rules, llm, hybrid) |
| ai_context_detected | string | no | Detected AI context (e.g., code_generation, image_generation) |
| is_public | boolean | no | Whether the template is public |
| tags | array | no | Tags for the template |
🔍 Retrieve a specific template by its ID.
| Parameter | Type | Required | Notes |
|---|---|---|---|
| template_id | string | yes | The ID of the template to retrieve |
✏️ Update an existing optimization template.
| Parameter | Type | Required | Notes |
|---|---|---|---|
| template_id | string | yes | The ID of the template to update |
| title | string | no | New title for the template |
| description | string | no | New description for the template |
| original_prompt | string | no | New original prompt text |
| optimized_prompt | string | no | New optimized prompt text |
| optimization_goals | array | no | New optimization goals |
| confidence_score | number | no | New confidence score (0.0-1.0) |
| model_used | string | no | New model used for optimization |
| optimization_tier | string | no | New tier of optimization |
| ai_context_detected | string | no | New detected AI context |
| is_public | boolean | no | Whether the template is public |
| tags | array | no | New tags for the template |
🗑️ Delete a saved optimization template by ID.
| Parameter | Type | Required | Notes |
|---|---|---|---|
| template_id | string | yes | The ID of the template to delete |
🔍 Search your saved template library with AI-aware filtering, context-based search, and sophisticated template matching
| Parameter | Type | Required | Notes |
|---|---|---|---|
| query | string | no | Search term to filter templates by content or title |
| ai_context | string | no | Filter templates by AI context type |
| sophistication_level | string | no | Filter by template sophistication level |
| complexity_level | string | no | Filter by template complexity level |
| optimization_strategy | string | no | Filter by optimization strategy used |
| limit | number | no | Number of templates to return (1-20) |
| page | number | no | Page number for pagination (use with limit to access results beyond the first page) |
| sort_by | string | no | Sort templates by field |
| sort_order | string | no | Sort order |
📋 List your most recently saved optimization templates, sorted by creation date.
| Parameter | Type | Required | Notes |
|---|---|---|---|
| limit | number | no | Number of recent templates to return (1-20) |
🧠 Detects the AI context for a given prompt using advanced backend analysis.
| Parameter | Type | Required | Notes |
|---|---|---|---|
| prompt | string | yes | The prompt text for which to detect the AI context |
Generate a structured SOP document for an AI agent from a goal description.
| Parameter | Type | Required | Notes |
|---|---|---|---|
| goal | string | yes | What the agent should accomplish |
| context | string | no | Additional context (optional) |
| model_id | string | no | Model to use (optional) |
| intent_frame | object | no | Optional IntentFrame to sharpen SOP scope and success criteria. |
Generate a complete skill package (SOP + SKILL.md + examples + helper.py) for an AI agent. Takes 30-120 seconds (async).
| Parameter | Type | Required | Notes |
|---|---|---|---|
| goal | string | yes | What the agent should accomplish |
| format | string | no | Output format |
| model_id | string | no | Model to use (optional) |
Transform a SOP into native code for LangChain, AutoGen, or Claude Code.
| Parameter | Type | Required | Notes |
|---|---|---|---|
| sop_content | string | yes | SOP content to transform |
| goal | string | yes | What the agent should accomplish |
| framework | string | yes | Target framework |
Check your Context Engineer credit balance and available workflow types.
No parameter schema on file.
Free, deterministic completeness check for an SOP or agent prompt (no LLM, no credits). Returns a 0-100 score and the missing sections.
| Parameter | Type | Required | Notes |
|---|---|---|---|
| sop_content | string | yes | SOP or agent prompt text (max 50,000 characters) |
| rubric_id | string | no | Rubric to score against (default: base) |
| Parameter | Type | Required | Notes |
|---|---|---|---|
| goal | string | yes | The workflow goal the harness is built for. |
| deploy_target | no | Target deployment platform(s). Single string or array (an array of more than one target requires Pro). All targets except claude_code and ollama require Pro. Default: claude_code. | |
| session_id | string | no | Optional: session ID from a prior generate_skill_package call to reuse SOP. |
| sop_content | string | no | The SOP content to base the harness on (required if no session_id). |
| output_dir | string | no | Directory to save the ZIP in. Default: OPTIMIZER_OUTPUT_DIR or a temp folder. |
| agent_read_only | boolean | no | Narrow the generated Claude Code subagent's tools to Read/Grep/Glob only. Set this for audit/review workflows that should never edit or execute anything. Default: false (full capability). |
| agent_harness | string | no | Execution backend for the generated agent.yaml: 'claude-sdk' (needs ANTHROPIC_API_KEY), 'codex' (needs OPENROUTER_API_KEY or OPENAI_API_KEY), or 'pi' (needs PI_API_KEY). Set this to match the API key |
| Parameter | Type | Required | Notes |
|---|---|---|---|
| goal | string | yes | The workflow goal to generate SOP variants for |
| context | string | no | Optional background context or documentation excerpt |
| blend_description | string | no | Optional: if provided, skips variant comparison and blends all 3 into one SOP using this description |
| perspective | string | no | Agent role or viewpoint (IntentFrame) |
| out_of_scope | string | no | What is explicitly excluded (IntentFrame) |
| success_definition | string | no | Measurable success criteria (IntentFrame) |
Fetch your latest published prompt template by slug for runtime use — decouple prompts from deploys.
| Parameter | Type | Required | Notes |
|---|---|---|---|
| slug | string | yes | The URL-safe slug of the prompt template (e.g., product-writer-a3f9c21b) |
Compile a prompt template with variable interpolation for runtime delivery — returns the fully interpolated prompt string ready for use.
| Parameter | Type | Required | Notes |
|---|---|---|---|
| slug | string | yes | The URL-safe slug of the prompt template |
| variables | object | no | Variable values to interpolate (e.g. {"user_name": "Alex", "plan": "Pro"}) |
List all version snapshots of a saved template — every update creates a snapshot you can inspect or restore.
| Parameter | Type | Required | Notes |
|---|---|---|---|
| template_id | string | yes | The ID of the template |
Restore a template to a previous version snapshot — undo unwanted changes instantly.
| Parameter | Type | Required | Notes |
|---|---|---|---|
| template_id | string | yes | The ID of the template |
| version_number | number | yes | The version number to roll back to (use list_template_versions to find available versions) |
Publish a template — makes it available for runtime delivery via get_prompt_by_slug.
| Parameter | Type | Required | Notes |
|---|---|---|---|
| template_id | string | yes | The ID of the template to publish |
Run a stateless one-shot evaluation of an optimized prompt using LLM judges — get actionable quality scoring without creating a dataset.
| Parameter | Type | Required | Notes |
|---|---|---|---|
| prompt | string | yes | The optimized prompt to evaluate |
| original_prompt | string | yes | The original prompt for comparison scoring |
🧠 Experimental: Bayesian optimization insights. The hosted service does not currently provide them, so this reports that insights are unavailable. Off by default (set ENABLE_BAYESIAN_OPTIMIZATION=true to list it)
| Parameter | Type | Required | Notes |
|---|---|---|---|
| analysis_depth | string | no | Depth of analysis to provide |
| include_recommendations | boolean | no | Include optimization recommendations |
⚡ Get real-time optimization status, AG-UI capabilities, and streaming optimization availability
No parameter schema on file.
- Every tool, no call limit
- No card, and mcprush charges nothing for it
- Source is public; its licence metadata reads NOASSERTION rather than naming a licence
- Runs on your machine — nothing of it reaches our gateway
- Nothing to cap, because mcprush bills nothing
What counts against your monthly calls
| Tool | Unit | Calls used | Out of the allowance |
|---|
Nothing here is billable: mcprush charges nothing to install Prompt Optimizer or to call it, at any volume.
Two independent axes, because powerful and malicious are different questions. The grade is threat only. The capability level is blast radius, and it is never a penalty on the grade — it is priced as one subtract-only term in the score, where you can see it.
| Term | Level | What it prices | Points |
|---|---|---|---|
| capability-exposure | minimal | capability blast radius (minimal) — client exposure if the model is manipulated | −0 |
| verification-discount | repo | publisher verification (public source) — no provenance, but the source is public and inspectable | −1 |
| coverage-honesty | source | inspection depth (source) — how much of the target the scan could see | −0 |
What the scan could actually read
A grade is only as meaningful as its coverage, so the scanner publishes its own depth before it publishes its result.
Tools were statically extracted from the published source (24 recovered), not enumerated from a running server. Tool-poisoning, Unicode-smuggling, capability and toxic-flow analysis ran on this inferred surface, but a mis-parsed registration could be missed or mis-attributed, so tool-derived findings are capped below “confirmed”. To grade the real runtime surface, scan the running server: --command "npx -y <package>".
1 of the 24 recovered tools are registered only when this.bayesianOptimizationEnabled holds (index.js:628) — typically when a key or option is configured. They are part of the server's surface and were analysed; a server started without that configuration exposes only the other 22.
1 of the 24 recovered tools are registered only when this.aguiFeatures holds (index.js:652) — typically when a key or option is configured. They are part of the server's surface and were analysed; a server started without that configuration exposes only the other 22.
Capability — what it could do if the model were manipulated
Tags derived from each tool’s schema and the implementation, not from what the tool calls itself. minimal is the level these add up to.
22 tools are registered unconditionally · (1) 1 more only when this.bayesianOptimizationEnabled holds (index.js:628) · (2) 1 more only when this.aguiFeatures holds (index.js:652) — typically when a key or option is configured. All 24 are in the table below and were analysed; the conditional ones are marked.
| Tool | Capability tags | Why the tag was assigned |
|---|---|---|
| optimize_prompt | no tags | |
| get_quota_status | no tags | |
| create_template | no tags | |
| get_template | no tags | |
| update_template | no tags | |
| delete_template | no tags | |
| search_templates | no tags | |
| list_recent_templates | no tags | |
| detect_ai_context | no tags | |
| generate_agent_sop | no tags | |
| generate_skill_package | no tags | |
| transform_for_framework | no tags | |
| get_ce_quota_status | no tags | |
| score_sop | no tags | |
| generate_harness_bundle | no tags | |
| explore_sop_approaches | no tags | |
| get_prompt_by_slug | untrusted-input | untrusted-input: keyword "fetch" |
| compile_prompt | no tags | |
| list_template_versions | no tags | |
| rollback_template | no tags | |
| publish_template | no tags | |
| run_quick_evaluation | no tags | |
| get_optimization_insights conditional (1) | no tags | |
| get_real_time_status conditional (2) | no tags |
Toxic-flow graph
The lethal trifecta, checked as a graph rather than as a checklist: untrusted input, a sensitive source and an external sink have to meet before there is a path worth worrying about.
The public result for this release does not print the flow graph, so there is nothing to show here. That is not the same as "no paths were found": what the scan did read is above, under coverage.
Supply chain and provenance
This is the first scan of this surface here, so there is nothing yet to compare it against.
Every result on this tab comes from one deterministic pass over the published package — offline, rule by rule, and auditable line by line above. Same methodology version, same bytes, same score.
Release history
Always latest — the install command names no version, so it installs the newest release when it runs; the current release on file is 3.8.11. Pin a release below and the command asks for exactly that one.
No release note was published with this version.
One review per account, from a signed-in account that does not publish this listing. Nothing else is asked — you can say what you think before you install it. Publishers can reply once per review, and a review can be edited or deleted by whoever wrote it.
Any signed-in account that does not publish Prompt Optimizer can review it, once — before installing it or after. A review from an account that has installed it is marked as one.
Nobody has reviewed this listing. The rating on the card is the mean of the reviews written here and nothing else, so there is no rating until somebody writes the first — which takes a signed-in account that does not publish it, and nothing else.