Promptzilla
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| PROMPTZILLA_RUNS_DIR | No | Override the directory for pipeline run storage. Default is ~/.promptzilla-mcp/runs. | ~/.promptzilla-mcp/runs |
| PROMPTZILLA_SKILLS_DIR | No | Override the bundled skills directory. Default is the packaged skills/ directory. | Bundled skills/ inside the package |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": true
} |
| prompts | {
"listChanged": true
} |
| resources | {
"listChanged": true
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| list_skillsB | List all available prompt-engineering skills. |
| search_skillsB | Keyword-search skill names and descriptions. |
| get_skillC | Load exact skill instructions and metadata for host execution. |
| taskerC | Prepare Tasker to optimize a rough task assignment. Host executes the returned skill; underlying task is never performed. |
| metapodC | Prepare Metapod to turn a raw request into System and Task Prompts. Host executes the returned packet. |
| judge_dreddB | Prepare a fresh Judge Dredd evaluation. Returns skill instructions plus caller-supplied prompt data, not a computed score. |
| brian_spillnerC | Prepare Brian Spillner to rewrite a prompt using a numbered scorecard. Host executes the packet. |
| prompt_pipelineA | Start a durable Full (idea → prompt) or optimize-only (existing prompt) run. Returns the next isolated-worker dispatch; the host executes it. |
| pipeline_statusB | Read run status, score history, artifacts and the active dispatch. |
| pipeline_submitA | Submit exactly one isolated worker output. Validate and persist it, then advance or retry the expected stage. |
| pipeline_pauseA | Pause a running stage for a real user decision or unavailable fresh workers. Preserve the pending stage for resume. |
| pipeline_resumeB | Resume a paused run using the actual user decision. Never invent an answer or bypass the four-round cap. |
| pipeline_apply_fixesA | Apply explicit Judge-authorized exact replacements to a STOP final. Preserves original final; requires fresh diff-only confirmation before approval. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
| brian-spillner | Rewrites a prompt by applying every numbered recommendation from a scorecard, cutting nothing that was not asked for, and returns the complete rewrite between BODY markers followed by a Rec-application checklist. Use when a scored prompt needs a revision that implements specific recommendations. |
| judge-dredd | Scores a system prompt on a locked 0-10 rubric and returns a scorecard with quoted evidence, an overall rating, a STOP or CONTINUE verdict and numbered recommendations. Use when a prompt needs an independent evaluation, when comparing prompts, or when re-scoring a rewrite. It evaluates only; it never rewrites the prompt or does the prompt's task. |
| metapod | Turns a raw request or idea into two prompts, a System Prompt and a Task Prompt, and never does the requested task itself. Use when someone wants a new assistant or agent prompt written from an informal request. |
| prompt-pipeline | Runs a prompt through Metapod (generate), Judge Dredd (score) and Brian Spillner (rewrite) until Judge says STOP at an overall score of 9 or higher. Use when the user wants a new prompt built and polished, or wants an existing prompt scored and improved. |
| tasker | This skill should be used when the user hands over a rough, free-form task assignment meant for another AI agent and wants it optimized into an execution-ready instruction without the task itself being performed. Use when the user says "optimize this task", "rewrite this instruction for an agent", "make this prompt better", "clean up this task assignment", "tasker", or pastes a rough delegation and asks for it to be turned into a clear, complete agent instruction. |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
| catalog | Discover skills and supporting assets. |
| brian-spillner | Rewrites a prompt by applying every numbered recommendation from a scorecard, cutting nothing that was not asked for, and returns the complete rewrite between BODY markers followed by a Rec-application checklist. Use when a scored prompt needs a revision that implements specific recommendations. |
| judge-dredd | Scores a system prompt on a locked 0-10 rubric and returns a scorecard with quoted evidence, an overall rating, a STOP or CONTINUE verdict and numbered recommendations. Use when a prompt needs an independent evaluation, when comparing prompts, or when re-scoring a rewrite. It evaluates only; it never rewrites the prompt or does the prompt's task. |
| metapod | Turns a raw request or idea into two prompts, a System Prompt and a Task Prompt, and never does the requested task itself. Use when someone wants a new assistant or agent prompt written from an informal request. |
| prompt-pipeline | Runs a prompt through Metapod (generate), Judge Dredd (score) and Brian Spillner (rewrite) until Judge says STOP at an overall score of 9 or higher. Use when the user wants a new prompt built and polished, or wants an existing prompt scored and improved. |
| tasker | This skill should be used when the user hands over a rough, free-form task assignment meant for another AI agent and wants it optimized into an execution-ready instruction without the task itself being performed. Use when the user says "optimize this task", "rewrite this instruction for an agent", "make this prompt better", "clean up this task assignment", "tasker", or pastes a rough delegation and asks for it to be turned into a clear, complete agent instruction. |
| asset:VERIFICATION.md | Original supporting file: VERIFICATION.md; read-only, never executed. |
| asset:brian-spillner/SKILL.md | Original supporting file: brian-spillner/SKILL.md; read-only, never executed. |
| asset:judge-dredd/SKILL.md | Original supporting file: judge-dredd/SKILL.md; read-only, never executed. |
| asset:metapod/SKILL.md | Original supporting file: metapod/SKILL.md; read-only, never executed. |
| asset:prompt-pipeline/DRY-RUN.md | Original supporting file: prompt-pipeline/DRY-RUN.md; read-only, never executed. |
| asset:prompt-pipeline/SKILL.md | Original supporting file: prompt-pipeline/SKILL.md; read-only, never executed. |
| asset:prompt-pipeline/scripts/exact_replace.py | Original supporting file: prompt-pipeline/scripts/exact_replace.py; read-only, never executed. |
| asset:prompt-pipeline/scripts/stage_control.py | Original supporting file: prompt-pipeline/scripts/stage_control.py; read-only, never executed. |
| asset:tasker/SKILL.md | Original supporting file: tasker/SKILL.md; read-only, never executed. |
TDQS
Scored across 13 tools
The pipeline_* tools are clearly differentiated by distinct verbs (status, submit, pause, resume, apply_fixes), and the skill discovery trio (list/search/get) is distinct. However, the four persona tools (tasker, metapod, judge_dredd, brian_spillner) share an identical 'prepare X to return skill instructions' behavior, and get_skill also returns skill instructions, so boundaries must be inferred from descriptions rather than names.
Two coherent families exist: verb_noun snake_case (list_skills, search_skills, get_skill) and noun-prefixed pipeline_* operations. But the four single-word persona names (tasker, metapod, judge_dredd, brian_spillner) break any verb_noun pattern, making the overall set mixed rather than predictable.
13 tools is well within a reasonable range for a prompt-engineering pipeline with skill discovery, skill wrappers, and a full run lifecycle. Each tool appears to earn its place, though the four persona wrappers could arguably be consolidated into a single parameterized tool.
The surface covers discovery (list/search/get), execution packaging (persona tools), and a durable pipeline lifecycle (start, status, submit, pause, resume, apply_fixes). Minor gaps exist—no explicit cancel/delete for runs or a listing of past runs—but core workflows are covered without obvious dead ends.