Skip to main content
Glama

Colour Memory

by app.railway.up.colour-memory-api-production

Generate Colour Direction for Another AI

agent_brief
Read-only

Generate a complete colour direction package for another AI agent or image generation model. Fetches a historically grounded archive palette from the concept, then produces: an agent brief (colour direction in prose), colour tokens with hex values and roles, a model-specific image generation prompt, a negative prompt, and lighting notes. Supports midjourney, flux, dalle, stable_diffusion. Example: task='luxury hotel bedroom', concept='Ottoman winter luxury', model='midjourney'. Use this to make Colour Memory the colour layer for other AI systems. Archive-grounded retrieval is evidence-filtered: entries with claim_role='reject' (no primary source and no period connection), stub entries, blank-source entries, and entries below minimum_claim_strength are never selected. If fewer than palette_size colours pass these filters, the call returns an honest incomplete result (ok:false, error_code:INSUFFICIENT_EVIDENCE_ELIGIBLE_PALETTE) rather than padding the palette with rejected or weak evidence. Negative constraints (from 'avoid' or negation phrases in concept like 'must never', 'without', 'not') are also applied to retrieval, not just flagged afterward -- a brief that says a wedding must never feel funereal will not surface mourning-themed colours in the first place. locked_palette calls skip evidence filtering entirely since the caller is supplying colours directly, not requesting archive evidence.

Input Schema

TableJSON Schema
NameRequiredDescriptionDefault
taskYesWhat the other AI needs to generate e.g. 'luxury hotel bedroom image'
avoidNoOptional: terms/associations to keep out of retrieval e.g. ['mourning', 'funeral black']. Affects retrieval directly, not just post-hoc flagging.
modelNoTarget model: midjourney, flux, dalle, stable_diffusionmidjourney
archiveNoOptional: restrict palette query to this archive e.g. georgianpleasures, japan, china
conceptYesColour concept to draw from e.g. 'Ottoman winter luxury', 'Victorian mourning'. Negation phrases ('must never', 'without', 'not X') are extracted and kept out of retrieval.
style_notesNoOptional: additional style direction e.g. 'matte surfaces only', 'no gold'
palette_sizeNoNumber of archive colours to include (default 5, max 8)
locked_paletteNoOptional: list of hex values to use exclusively. When provided, no archive query is run and no evidence filtering applies — these exact colours are used as-is. Prevents palette drift.
allowed_archivesNoOptional: list of allowed archive names. Query restricted to these archives only.
minimum_claim_strengthNoEvidence-grade floor for archive-grounded retrieval (default 'C'). Grade controls documentary strength; claim_role='reject' entries are excluded regardless of how low this is set. Not applied to locked_palette calls.C

Output Schema

TableJSON Schema
NameRequiredDescriptionDefault

No arguments

TDQS

A4.6/5.0
Behavior5/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

ReadOnlyHint already says no mutation; the description adds substantial behavioral detail: evidence filtering rules, honest incomplete results (ok:false, INSUFFICIENT_EVIDENCE_ELIGIBLE_PALETTE), negative constraints applied at retrieval time, and locked_palette skipping all filtering. This far exceeds the annotation baseline.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is long and organized in paragraphs, but every section adds necessary behavior context for a 10-param tool. It is front-loaded with a clear summary sentence, then details filters, error handling, and special modes. No filler, though the length keeps it from a 5.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness5/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Given the output schema exists, the description doesn't need to list return fields, but it still enumerates the package components (brief, tokens, prompt, negative prompt, lighting notes), covers error codes and edge cases, and gives supported models. This is complete for the tool's complexity.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

All 10 parameters have schema descriptions (100% coverage), so baseline is 3. The description adds meaning beyond schema by explaining how avoid and negation phrases affect retrieval, what minimum_claim_strength means in practice (reject entries excluded regardless), and that locked_palette uses colours as-is. It doesn't cover every parameter but the critical ones get enriched, so 4.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The first sentence states a specific verb+resource: 'Generate a complete colour direction package for another AI agent or image generation model.' It explicitly names the deliverable and distinguishes from sibling palette/image tools by the 'for another AI' framing. Example task/concept/model anchors real usage.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines4/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description says 'Use this to make Colour Memory the colour layer for other AI systems,' providing clear when-to-use context. It also contrasts locked_palette vs archive-grounded retrieval paths, but it does not explicitly name alternatives among siblings or give when-not-to-use exclusions, so a 4.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

All Colour Memory toolsInspect
Try in Browser
RayrunBuild, deploy, and operate MCP servers without becoming an infrastructure team.ray.runAd

Glama MCP Gateway

Add one secure layer between your agents and this server.

TDQS

A3.7/5.0
Disambiguation2/5

Multiple tool clusters have unclear boundaries: colour_passport, colour_dna, colour_metrics, and colour_cultural_risk overlap heavily; palette_extract, extract_image_colours, image_palette, and ingest_image cover similar image-colour extraction territory; and compound tools like design_session, image_brief, session_brief, and archive_report_brief duplicate chains of component tools. The descriptions are detailed and occasionally state 'use this instead of X', but an agent still faces many near-duplicate choices.

Naming Consistency3/5

Names are consistently snake_case and often use domain prefixes (colour_, palette_, archive_, brand_, accessibility_), but verb placement is mixed: some are verb_noun (extract_image_colours, query_hex, style_match), others are noun_verb (colour_dna, palette_generate, brand_audit), and a few standalone names (ui_states, tool_guide, meta_capabilities) don't fit either pattern. The convention is readable but not uniform.

Tool Count1/5

At 88 tools, this is far beyond the reasonable well-scoped range and exceeds the 50+ extreme mismatch threshold. Many tools are compound wrappers that consolidate chains of simpler tools, adding redundancy and cognitive load rather than genuine coverage.

Completeness5/5

For the apparent domain, the tool surface is extremely comprehensive: colour extraction, analysis, naming, accessibility, cultural/provenance research, palette generation, brand systems, interior design, ecommerce copy, image briefs, project lifecycle, exports, and diagnostic tools are all present. Persistent objects have list/get/versions/delete/export support, so there are no obvious dead ends.

Resources