Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add neul-labs/brat --skill brat-architecturegit clone --depth 1 https://github.com/neul-labs/bratWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/neul-labs/brat/brat-architecture)<a href="https://agentmods.dev/skills/neul-labs/brat/brat-architecture"><img src="https://agentmods.dev/badge/skills/neul-labs/brat/brat-architecture/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/neul-labs/brat/brat-architecture"><img src="https://agentmods.dev/badge/skills/neul-labs/brat/brat-architecture.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00027 | $0.00527 |
| Opus 5 | $0.00014 | $0.00264 |
| Sonnet 5 | $0.00005 | $0.00105 |
| Haiku 4.5 | $0.00003 | $0.00053 |
Grade A, and why
brat-architecture scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 10d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 76 lines — stays where its author put it; the contents beside it link to each section on GitHub.
brat-architecture Skill
Purpose
Read product requirements and produce architecture design decisions, component breakdown, and interface definitions.
Input Schema (JSON)
{
"product_requirements": {
"title": "string",
"user_stories": [{"story": "string", "acceptance_criteria": ["string"]}],
"acceptance_tests": ["string"]
},
"architecture_context": "string (relevant arch notes from KB)",
"existing_components": ["string (known components from KB)"],
"tech_stack": "string (language/framework info)"
}
Output Schema (JSON)
{
"action": "create_design|update_design|answer",
"design": {
"title": "string",
"components": [
{
"name": "string",
"responsibility": "string",
"interfaces": ["string (input/output contracts)"],
"dependencies": ["string (other components)"],
"file_paths": ["string (suggested files/modules)"]
}
],
"adr": {
"title": "string",
"context": "string",
"decision": "string",
"consequences": ["string"]
},
"test_strategy": {
"unit_tests": ["string (what to test per component)"],
"integration_tests": ["string (cross-component tests)"]
},
"notes_to_create": [
{
"title": "string",
"body": "string",
"tags": ["architecture", "adr", "component"],
"type": "permanent|structure",
"links_to": ["string (existing note slugs)"]
}
]
},
"answer": "string (if action=answer)",
"escalation_reason": "string (if requirements are unclear)"
}
Guardrails
- ALWAYS read product requirements before designing
- Components must have clear single responsibilities
- Every component interface needs input/output contracts
- Create ADR (Architecture Decision Record) for significant choices
- Link architecture notes to product notes they implement
- Define test strategy BEFORE implementation (TDD)
- NEVER write implementation code in this phase
- NEVER proceed to implementation from this skill without explicit test strategy
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 10d ago First seen · 76 lines · 0 tokens per session scan A cc2f5c5e1425
brat-architecture is a skill published in the GitHub repository neul-labs/brat (5 stars, last pushed 2mo ago), licensed MIT. It adds 27 tokens to every session and 527 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
swarm-pr-review
Run a graph-guided, tool-augmented PR review using context packing, parallel exploration, mandatory repository-agnostic risk-family coverage with dispatch scaled to diff size and risk, independent reviewer validation, critic challenge, and metrics writeback. Use for deep pull request review with low false-positive…
workflow
Use when a task is too large for turn-by-turn orchestration and should run through the big-task workflow lane: system-wide changes, large migrations, repo-wide audits, high-confidence verification, or tasks explicitly asking to run a workflow. Claude Code uses native dynamic workflows; Codex, OpenCode, and Grok use…
swarm-plan
Full execution protocol for MODE: PLAN -- plan creation, external plan ingestion, QA gate persistence, task granularity, and traceability checks.
loop
Full execution protocol for MODE: LOOP — the compound-engineering loop: brainstorm → plan → build → review → improve, iterating under defense-in-depth stop conditions with generator/critic separation, durable resumable state, and mandatory compounding learning capture. Loaded on demand by the architect when the loop…
council
Full execution protocol for MODE: COUNCIL -- General Council research, parallel member dispatch, disagreement handling, and synthesis.
deep-research
Full execution protocol for MODE: DEEPRESEARCH — orchestrator-worker deep research over external sources: decompose, iterative websearch/webfetch retrieval, parallel sme synthesis, dual-reviewer claim verification, critic challenge of high-stakes claims, and a cited report. Loaded on demand by the architect when the…