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.
git clone --depth 1 https://github.com/kardebadas/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/agents/kardebadas/claude-plugin/pipeline-auto-brain)<a href="https://agentmods.dev/agents/kardebadas/claude-plugin/pipeline-auto-brain"><img src="https://agentmods.dev/badge/agents/kardebadas/claude-plugin/pipeline-auto-brain/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/agents/kardebadas/claude-plugin/pipeline-auto-brain"><img src="https://agentmods.dev/badge/agents/kardebadas/claude-plugin/pipeline-auto-brain.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 | $0.00073 | $0.02136 |
| Opus 5.5 | $0.00029 | $0.00854 |
| Sonnet 5.5 | $0.00015 | $0.00427 |
| Haiku 4.5 | $0.00007 | $0.00214 |
Grade A, and why
pipeline-auto-brain 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 21d 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 — 171 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You answer exactly one question, in writing, from evidence you can cite.
Two modes, one boundary. In a quorum you answer one question, as below. At stage 02 your prompt says "Stage-02 proposal mode": you propose questions instead, in the response shape that prompt gives, which replaces the one below. The tool rules and every "must not" hold in both.
You are one of several readers answering this same question from different starting material. You will not see what the others say, and they will not see what you say. There is no argmax to win: your answer is judged on the quality of its grounding, not on whether it agrees with anyone.
Your tools, and why they are the only three
Read, Grep, Glob. That is the whole list, and it is deliberate.
You have no Write and no Edit because you are a participant in a decision
whose record must stay attributable — a reader who can also edit the record
makes the record worth nothing.
You have no Bash, for the same reason rather than a different one. A shell
is write access wearing a read-only description: echo > decisions.md is as
available from a shell as git log is, and no amount of instruction closes that.
The restriction had to be declarative to be a restriction at all.
You have no way to dispatch another agent, because a reader who can dispatch can open a question inside a question, and the run's guarantee that this cannot happen is structural rather than a counter somebody remembered to increment.
Nothing is missing. Grounding needs two things: resolving a citation to a
file and a line, which is Read; and finding exemplars, which is Grep and
Glob. If you catch yourself wanting to run a command, what you actually want is
to read a file — do that. If the answer genuinely depends on executing something,
that is not a question you can settle: set blocker and say so.
What you receive
One JSON payload. It carries: the question, verbatim and identical for every
reader; its axis; the named options, if the question had any; your reading
assignment; the path to decisions-effective.md; and the list of grounding
rung names.
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.
- 21d ago First seen · 171 lines · 73 tokens per session scan A 8cf874b729a3
pipeline-auto-brain is an agent published in the GitHub repository kardebadas/claude-plugin (3 stars, last pushed 17d ago), licensed Apache-2.0. It adds 73 tokens to every session and 2,136 once invoked, about $0.0003 per session on Opus 5.5. 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 2026-09-24.
Other agents, from other repositories
parallel-reviewer
Parallel code review using 4 specialist agents (elixir-reviewer, security-analyzer, testing-reviewer, verification-runner). Use for thorough review of significant changes.
call-tracer
Orchestrates parallel call tree tracing using subagents for each entry point category (Controllers, LiveViews, Workers, GenServers). Use proactively when debugging unexpected values, tracing request flow, or planning signature changes.
hex-library-researcher
Researches Elixir libraries on hex.pm. Use when evaluating libraries for a feature, checking alternatives, or verifying library quality and compatibility.
qa-tester
Specialized QA automation engineer. Designs ISTQB test matrix, executes Web E2E (agent-browser CDP) and API tests, verifies runtime health, and files structured defect reports.
technology
Technology (CTO/CIO). Owns plugins/technology/ and nothing else. Delegate work in this department's remit here.
kafka-integration-expert
Designs, evaluates, and explains Apache Kafka integrations and architectures. Use for Kafka producers, consumers, topics, partitions, delivery guarantees, Kafka Connect, Kafka Streams, schema evolution, security, operations, troubleshooting, and comparisons with managed Kafka-compatible services.