ap-aesthetics
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
No arguments | |||
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 | {} |
| resources | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| ap_catalogA | List cue definitions, cognitive states, emotion formulas, parameters, audience profile and target-score meanings. |
| ap_templateC | Create a work skeleton with unknown cues. Stage descriptions are reference data, not instructions. |
| ap_evaluateC | Compute annotated media trajectories and explain weak spots. Does not infer media cues or claim population accuracy. |
| ap_compareB | Compare two creative artifacts with the same audience, target and formulas by default. Returns tradeoffs; higher score is not satisfaction probability. |
| ap_audience_panelB | Evaluate explicitly stated audience profiles and a weighted scenario mixture. Not a measured population sample. |
| ap_sensitivityC | Vary one parameter over explicit values to expose model sensitivity. Not a creative improvement. |
| ap_inspect_datasetB | Inspect observation provenance, missing labels and group/split leakage before calibration. |
| ap_fit_calibrationB | Fit a versioned local output calibrator using train/validation; test is held out. Does not adopt or overwrite any model. |
| ap_apply_calibrationB | Apply an explicit frozen calibration model to independent prediction channels. |
| ap_tune_parametersA | Evaluate an explicit finite parameter grid against independent human labels. Fit/select on train/validation, report test once, return candidate without adoption. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
| AP cognitive and emotion catalog |
TDQS
Scored across 10 tools
Each tool has a distinct verb+object focus: evaluate vs compare vs sensitivity vs audience_panel are separable evaluation modes, and fit/apply/tune form distinguishable calibration steps. Mild overlap among the evaluation tools (ap_evaluate, ap_compare, ap_sensitivity) is resolved by explicit descriptions.
All names share a consistent ap_ prefix and are snake_case, but the pattern mixes noun forms (ap_catalog, ap_template, ap_audience_panel, ap_sensitivity) with verb-based forms (ap_evaluate, ap_fit_calibration, ap_tune_parameters). Readable and predictable overall, just not uniformly verb_noun.
Ten tools is well within the ideal 3-15 range and each maps to a distinct capability in the evaluation/calibration pipeline. No redundant or filler tools are apparent.
The surface covers a full workflow: reference lookup, templating, evaluation, comparison, panel, sensitivity, dataset inspection, and fit/apply/tune calibration. Minor gaps around persistence/export of results, but core lifecycle is intact.