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 agentmods add skills/cbdreamer11/cb-loop-kit-claude-plugin/loop-councilnpx skills add cbdreamer11/CB-loop-kit-claude-plugin --skill loop-councilgit clone --depth 1 https://github.com/cbdreamer11/CB-loop-kit-claude-pluginWrote 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/cbdreamer11/cb-loop-kit-claude-plugin/loop-council)<a href="https://agentmods.dev/skills/cbdreamer11/cb-loop-kit-claude-plugin/loop-council"><img src="https://agentmods.dev/badge/skills/cbdreamer11/cb-loop-kit-claude-plugin/loop-council.svg" alt="Measured on agentmods" 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 | $0.00052 | $0.00525 |
| Opus 5 | $0.00026 | $0.00262 |
| Sonnet 5 | $0.00010 | $0.00105 |
| Haiku 4.5 | $0.00005 | $0.00052 |
Grade A, and why
loop-council 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 5d 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 — 51 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Council
A council exists so that a decision is made against reality instead of against recollection. It runs before the code, not as a review afterwards.
When it is required (not optional)
Look at what the change will touch — not at how confident you feel, because self-reported confidence is exactly the thing that fails:
- money, pricing, refunds, anything that charges someone
- permissions, authentication, or who can see whose data
- database schema, migrations, or anything that is hard to undo
- a public contract other code depends on
Two voices minimum. Three when it touches money, permissions, or schema.
The voices
Dispatch them in one message so they run in parallel, each blind to the others:
loop-planner— the architecture: the shape, the alternatives, what breaks later.loop-domain— the facts: decides against real documentation and the real installed behaviour, with URLs or exact strings. Never from memory.loop-auditor— the red team: its only job is to refute the proposed design before it exists. Give it the proposal, and tell it to assume there is a hole.
Resolving it
- Have each voice write its verdict to
.loop/council/<topic>-<voice>.mdas it finishes. Councils get interrupted; a verdict only in context is a verdict lost, and you should be able to resume without re-running the ones that finished. - If the red team and the architect disagree, the red team wins by default — or it escalates to the owner if it is really a business call. Never split the difference into something neither voice endorsed.
- Write the outcome into
.loop/DECISIONS.md: the decision, the alternative that was rejected and why, and the evidence. The rejected alternative is the part future sessions will need most.
What a council is not
It is not a vote to feel better about a decision already made, and it is not a substitute for reading the code. If the question can be answered by opening a file, open the file.
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.
- 5d ago First seen · 51 lines · 52 tokens per session scan A 71339b4483ff
loop-council is a skill published in the GitHub repository cbdreamer11/CB-loop-kit-claude-plugin (8 stars, last pushed 1mo ago), licensed MIT. It adds 52 tokens to every session and 525 once invoked, about $0.0003 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
orchestrator
Use when you are unsure which specialist to invoke, when a task spans multiple roles, when you want the system to plan before acting, or when you need parallel specialist agents dispatched to complete a complex task. The orchestrator breaks work into subtasks, assigns specialists, dispatches agents, and synthesises…
backend-engineer
Use when designing APIs, working with databases, building microservices, handling authentication and authorisation, optimising server performance, designing data models, or any task involving server-side logic, infrastructure, or system architecture.
kotlin-dev
Use when implementing any feature or bugfix in a Kotlin/Spring Boot service, covers layer conventions, idioms, testing, and observability.
data-analyst
Use when analysing data files (CSV, JSON, Excel), writing SQL queries, identifying trends and patterns, building dashboards or charts, summarising metrics, validating data quality, or turning raw data into actionable insights for decision-making.
devex
Use when improving developer workflows, setting up or optimising CI/CD pipelines, reducing build times, improving local development setup, evaluating developer tooling, writing internal documentation for engineers, measuring developer productivity, or any task focused on making the engineering team faster and less…
fintech-frontend-engineer
Use when building React applications or components in a fintech context, implementing financial dashboards, transaction histories, payment flows, or data-heavy UIs, working with Tailwind CSS, optimising for SEO (meta tags, structured data, Core Web Vitals), improving page load performance, handling sensitive financial…