Files

T

mpmedia 9e6d9043aa docs: update README to v1.2.1

2026-07-01 13:08:02 -05:00

12 KiB

Raw Blame History

mpm-brand-voice

Keeps every piece of AI-generated content aligned with Message Point Media's voice, tone, messaging pillars, and visual identity. Brand settings, the full logo catalog (PNG + SVG), and the Montserrat/Open Sans fonts are baked into the plugin payload — logos and fonts load instantly with no Google Drive or Google MCP fetch required. Works immediately in any working folder, no setup.

Version: 1.2.1 Author: Message Point Media Repo: https://git.mpm.to/mpm/mpm-brand-voice-plugin CoWork Project: CW-001 — MPM Brand Voice

Overview

The MPM Brand Voice plugin gives Claude a persistent, structured understanding of how Message Point Media communicates. It ships with a fully baked-in settings/brand-voice.local.md file containing MPM's complete brand identity — 11 messaging pillars, 6 audience personas, voice constants, tone matrix, product terminology, logo system, color palette, and typography references. As of v1.2.1, the plugin also bakes in the full MPM logo catalog (24 PNG/JPG + 40 SVG variants under assets/logos/) and the Montserrat/Open Sans variable font files (assets/fonts/), so logo and font lookups never require the Google Drive connector or Google MCP at runtime — the Drive folders remain the upstream source of truth but are no longer needed for day-to-day content generation. A working-folder override (.claude/brand-voice.local.md) can be placed in any project folder to supply project-specific positioning (e.g., a co-branded proposal) that takes precedence over the baked-in defaults. The plugin exposes two skills: one for applying the guidelines to content creation, and one for enriching the guidelines from new source materials.

Skills

Skill What It Does

brand-voice-enforcement Applies MPM brand guidelines to any AI-generated content. Loads brand settings via bash cat on settings/brand-voice.local.md (required — the Read tool silently fails on .remote-plugins paths; a failed read is a hard stop, never a silent fallback), identifies the content type and target audience persona, selects relevant messaging pillars, applies voice constants and tone flexes, resolves the correct baked-in logo file via the Logo Selection Rules table, generates content, and validates the output. Trigger phrases include: "write an email", "draft a proposal", "create a pitch deck", "write a LinkedIn post", "draft a presentation", "write a Slack message", "draft sales content", "on-brand", "brand voice", "enforce voice", "apply brand guidelines", "brand-aligned content", "write in our voice", "use our brand tone", "make this sound like MPM", "rewrite this in our tone", "this doesn't sound on-brand". Not for generating guidelines from scratch — use guideline-generation for that.

guideline-generation Generates or updates MPM brand voice guidelines from source materials — uploaded files (PDF, PPTX, DOCX, MD, TXT), discovery reports, working folder documents, Google Drive files, or direct user input. Always loads existing brand-voice.local.md before generating; builds on what exists rather than replacing it wholesale. Applies confidence scoring and source classification (Authoritative / Strong Signal / Supporting / Context Only). Trigger phrases include: "generate brand guidelines", "update our brand settings", "create a style guide", "extract brand voice", "analyze these materials for brand voice", "build guidelines from these documents", "synthesize a voice and tone guide", "update brand-voice.local.md", "add this content to our brand settings". Not for applying guidelines to content creation — use brand-voice-enforcement for that.

Skill	What It Does
`brand-voice-enforcement`	Applies MPM brand guidelines to any AI-generated content. Loads brand settings via `bash cat` on `settings/brand-voice.local.md` (required — the Read tool silently fails on `.remote-plugins` paths; a failed read is a hard stop, never a silent fallback), identifies the content type and target audience persona, selects relevant messaging pillars, applies voice constants and tone flexes, resolves the correct baked-in logo file via the Logo Selection Rules table, generates content, and validates the output. Trigger phrases include: "write an email", "draft a proposal", "create a pitch deck", "write a LinkedIn post", "draft a presentation", "write a Slack message", "draft sales content", "on-brand", "brand voice", "enforce voice", "apply brand guidelines", "brand-aligned content", "write in our voice", "use our brand tone", "make this sound like MPM", "rewrite this in our tone", "this doesn't sound on-brand". Not for generating guidelines from scratch — use `guideline-generation` for that.
`guideline-generation`	Generates or updates MPM brand voice guidelines from source materials — uploaded files (PDF, PPTX, DOCX, MD, TXT), discovery reports, working folder documents, Google Drive files, or direct user input. Always loads existing `brand-voice.local.md` before generating; builds on what exists rather than replacing it wholesale. Applies confidence scoring and source classification (Authoritative / Strong Signal / Supporting / Context Only). Trigger phrases include: "generate brand guidelines", "update our brand settings", "create a style guide", "extract brand voice", "analyze these materials for brand voice", "build guidelines from these documents", "synthesize a voice and tone guide", "update brand-voice.local.md", "add this content to our brand settings". Not for applying guidelines to content creation — use `brand-voice-enforcement` for that.

Tools Reference

This is a skill-only plugin — no MCP server, no external API calls. All processing is done by Claude using the baked-in skill prompts, settings, and asset files.

Skill	Reference Files
`brand-voice-enforcement`	`settings/brand-voice.local.md`, `references/voice-constant-tone-flexes.md`, `references/before-after-examples.md`, `references/mpm-brand-resources.md`, `../../assets/README.md`
`guideline-generation`	`settings/brand-voice.local.md`, `references/guideline-template.md`, `references/confidence-scoring.md`, `references/mpm-brand-resources.md`

Baked-In Assets (new in v1.2.1)

assets/logos/png/ — 24 raster logos (PNG/JPG) for documents, web, presentations, email. assets/logos/svg/ — 40 vector logos (SVG) for print and large-format. assets/fonts/ — Montserrat (headings) and Open Sans (body) variable fonts, plus their SIL Open Font License files. assets/README.md — full asset manifest: logo quick-pick table, directory layout, and the filename-normalization mapping back to the original Drive names (spaces/@/[] stripped for plugin-packaging safety).

Resolve any asset path relative to ${CLAUDE_PLUGIN_ROOT}, e.g. ${CLAUDE_PLUGIN_ROOT}/assets/logos/png/Logo-Standard-Dark.png. Always confirm a path with ls before handing it to a document tool.

Logo Selection Rules

Background	Logo to Use	File (in `assets/logos/png/`)
White or light backgrounds	Standard Dark (DEFAULT)	`Logo-Standard-Dark.png`
White / off-white backgrounds	Gold + Black text	`Logo-Gold-Black.png`
Dark, charcoal, or dark-gold backgrounds	Gold + White text	`Logo-Gold-White.png`
Cover pages, title blocks with dark header bar	Gold + White text	`Logo-Gold-White.png`
Single-color black print	Black text only	`Logo-Black-Text.png`
Dark background, icon only needed	White Cutout	`Logo-White-Cutout.png`

For print and large-format output, use the matching SVG from assets/logos/svg/. Transit-specific deliverables use the TransitPoint sub-brand lockup (MPM-Transitpoint-Dark.png / MPM-Transitpoint-Light.png).

Setup Instructions

No server to run, no environment variables required. The plugin is self-contained.

Step-by-step:

Obtain mpm-brand-voice-v1.2.1.plugin from the CW-001 Drive coordination folder
Open CoWork and navigate to Plugin settings
Click "Install Plugin" and select the .plugin file
Restart CoWork to activate
Verify by asking Claude to "write a short email in our voice" — it should reference MPM brand attributes, use a baked-in logo, and produce on-brand content without any additional setup

Setup Checklist

.plugin file obtained from CW-001 Drive folder
Plugin installed via CoWork Plugin settings
CoWork restarted
Test content generation confirms brand settings and baked-in logo/font assets loaded

Connection Details

Field	Value
Plugin type	Skill-only (no MCP server)
Transport	None (no external API)
Auth method	None required
Settings file	`settings/brand-voice.local.md` (baked-in)
Asset payload	`assets/logos/` (64 files), `assets/fonts/` (baked-in, no Drive fetch)
Working folder override	`.claude/brand-voice.local.md` (optional, project-specific)

Workflow Notes

Brand Settings Priority Order When a skill loads settings, it checks in this order and stops at the first match:

.claude/brand-voice.local.md in the current working folder (project-specific override)
settings/brand-voice.local.md baked into the plugin (always available)
Brand guidelines generated earlier in the session via guideline-generation

Voice vs. Tone Voice is WHO the brand is — it never changes. Tone is HOW it speaks in a given moment — it flexes along three dimensions: Formality (High/Medium/Low), Energy (High/Medium/Warm/Low), and Technical Depth (High/Medium/Low). The skill applies the correct tone matrix for the content type and audience automatically.

Content Type Guidance

Cold outreach: Medium formality, High energy, Low depth — hook fast, earn attention
Enterprise proposals: High formality, Medium energy, High depth — ROI and precision
Social media: Low-Medium formality, High energy, Low depth — brevity and personality
Customer success: Medium formality, Warm energy, Medium depth — empathy and competence

Conference Language "Born to Innovate. Built to Last." is the primary tagline for standard use. Motorsports/race-themed language (pit stop, velocity, etc.) is MPower 2025 conference-specific. The skill will flag use of this language outside that context and offer the standard brand equivalent.

Keeping Settings and Assets Current When the MPM brand evolves:

Update the master — edit .claude/brand-voice.local.md in the Claude CoWork working folder on Google Drive, or update the baked-in settings/brand-voice.local.md source directly for a plugin-wide change
If logos or fonts changed, re-pull the affected files from the MPM Logos Drive folder into assets/ and re-run the filename normalization in assets/README.md
Repackage — ask Claude to "repackage the MPM brand voice plugin with the updated settings"
Push to org library — upload the new .plugin file; team members will see an update available

Brand Document Locations (Google Drive)

Document	Location
MPM Brand Guide (.docx / .pdf)	Design Guide folder
MPM Logo Guide (.docx / .pdf)	Design Guide folder
MPM Brand Deck (.pptx)	Design Guide folder
Logo files (PNG / SVG) — upstream source of truth	MPM Logos → 01 New Logos
Fonts — upstream source (Google Fonts originals)	02 MPM Branding → Fonts
Zoom backgrounds, templates	02 MPM Branding

Design Guide: https://drive.google.com/drive/folders/1-H19hqC6HUX_3FjjrRVAoimg688CldrC Logos: https://drive.google.com/drive/folders/1xpO_sh0dQvWiMQbdmbhRZARuev4RpZiL

Requirements

CoWork (Claude Desktop with plugin support)
No external dependencies, API keys, or environment variables

Troubleshooting

Symptom	Fix
Content doesn't reflect MPM voice	Confirm plugin is installed and CoWork was restarted after install
Wrong brand settings loaded	Check for a `.claude/brand-voice.local.md` in the working folder — it overrides the plugin defaults
Guideline generation overwrites existing settings	The skill should enrich, not replace. If it replaced, restore from the `brand-voice.local-YYYY-MM-DD.md` backup the skill creates before saving
Conference/race language appearing in standard content	The skill should flag this — if it didn't, ensure you're using the `brand-voice-enforcement` skill, not a generic Claude response
Logo file can't be resolved	Confirm `${CLAUDE_PLUGIN_ROOT}` is set and run `ls` on `assets/logos/png/` to diagnose; never substitute a non-MPM logo — use the labeled placeholder pattern in `brand-voice-enforcement/SKILL.md` instead

Version History

Version	Date	Notes
1.2.1	2026-07-01	Full MPM logo catalog (64 files) and Montserrat/Open Sans variable fonts baked into `assets/` — eliminates Google Drive/Google MCP fetch for logos and fonts at runtime. New `assets/README.md` manifest with filename-normalization mapping. `SKILL.md` updated to resolve asset paths via `${CLAUDE_PLUGIN_ROOT}`. Keywords updated (added logos, fonts, offline-assets).
1.1.0	2026-05-21	`bash cat` fix for brand settings read (Read tool silently fails on `.remote-plugins` — hard stop on failure). Logo Usage section added to `SKILL.md` with selection table, placement rules, and fallback behavior. Visual Identity block added to Quick Reference with Montserrat/Open Sans requirements and full hex palette.
1.0.0	2026-05-05	Initial release. Full MPM brand settings including MPower 2025 conference content, complete logo catalog, font references, 11 messaging pillars, 6 audience personas, and product terminology.

12 KiB Raw Blame History