A style guide for prose written to and by agents: a cascade of always-on invariants, registers, and per-artifact modules, plus adversarial validators that keep the guide honest.
Attacks every TAMOS rule for ambiguity — unscoped quantifiers, dangling referents, vague terms an agent could misread two ways.
Audits the TAMOS enforced layer (core, registers, artifacts) for rationale, filler, or prose that should live in the reference layer instead.
Tests whether TAMOS artifact modules respect the cascade — stating only deltas, never re-litigating rules they inherit from a register or core.
Finds cross-tier conflicts in TAMOS — any Tier-2 rule that contradicts a rule it inherits from its register or core.
Checks that TAMOS obeys its own rules — every file, including docs, is held to the declarative register it preaches.
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
A style guide for prose written to agents and by agents.
Chicago tells one writer how to address one reader. TAMOS governs two directions at once: how humans write to agents (prompts, specs, instructions) and how agents write back (reviews, reports, commits, docs). Those are different registers with different failure modes, and most bad "agent writing" comes from conflating them.
A cascade. General rules at the top; specialized overrides below; the more specific rule wins on conflict.
Tier 0 — CORE universal invariants, both directions → core.md
Tier 1 — REGISTERS imperative (to agents) / declarative (from agents)
Tier 2 — ARTIFACTS per-output deltas (PR review, commit, report…)
A Tier-2 module states only its delta over the register it inherits. The PR-review module does not re-explain how to cite or hedge — it says only what is special about a PR review. ("Style" names the discipline as a whole; a single unit is always a "module".)
| Layer | Lives in | Loaded | Style |
|---|---|---|---|
| Enforced | CLAUDE.md / prompt | every turn | terse imperatives, no why |
| Reference | this repo | on demand | prose, rationale, examples |
Rationale never goes in the enforced layer. The why is for the human
maintaining the guide and lives in docs/.
TAMOS ships as a Claude Code plugin (the repo is its own marketplace):
/plugin marketplace add Battle-Creek-LLC/tamos
/plugin install tamos@battle-creek
Then opt the always-on layer into your global ~/.claude/CLAUDE.md by import.
Full steps, CI/CD, and release flow are in docs/user-guide.md.
docs/principles.md — the philosophy (two registers, token economy, dogfooding)docs/structure.md — the tier cascade, the fixed artifact shape, deliverydocs/runtime.md — how modules load (pull/skill) and get enforced (hook)docs/user-guide.md — install, always-on import, validation, CI/CD, releasesdocs/contributing.md — how to add a new artifact moduleAGENT-STYLE.md — the index/registry of every moduleagents/ — adversarial validators that attack this guide for defectsnpx claudepluginhub battle-creek-llc/tamos --plugin tamosComprehensive UI/UX design plugin for mobile (iOS, Android, React Native) and web applications with design systems, accessibility, and modern patterns
Multi-model consensus engine integrating OpenAI Codex CLI, Gemini CLI, and Claude CLI for collaborative code review and problem-solving.
Ultra-compressed communication mode. Cuts ~75% of tokens while keeping full technical accuracy by speaking like a caveman.
Standalone image generation plugin using Nano Banana MCP server. Generates and edits images, icons, diagrams, patterns, and visual assets via Gemini image models. No Gemini CLI dependency required.
Write feature specs, plan roadmaps, and synthesize user research faster. Keep stakeholders updated and stay ahead of the competitive landscape.
Curate auto-memory, promote learnings to CLAUDE.md and rules, extract proven patterns into reusable skills.