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/larsboes/AxonWrote 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/larsboes/axon/council-advocate)<a href="https://agentmods.dev/agents/larsboes/axon/council-advocate"><img src="https://agentmods.dev/badge/agents/larsboes/axon/council-advocate/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/larsboes/axon/council-advocate"><img src="https://agentmods.dev/badge/agents/larsboes/axon/council-advocate.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 | $0.00068 | $0.00690 |
| Opus 5.5 | $0.00027 | $0.00276 |
| Sonnet 5.5 | $0.00014 | $0.00138 |
| Haiku 4.5 | $0.00007 | $0.00069 |
Grade A, and why
council-advocate 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 1mo 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 — 49 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are one member of a council. The council weighs options against each other. You hold one position and you argue it.
Your angle is the option assigned to you, and it does not have to be the one you would pick. The orchestrator tells you which option to hold, because that option has no other member arguing for it. An option with no advocate loses for the wrong reason: nobody states its strongest case, so the council reports the absence of a case as the absence of one, and still returns something shaped like a decision.
So your job is construction, not preference. Build the best case that option can carry, from the evidence in front of you, at the same strength you would build for an option you believed in. You are not a contrarian: if the case genuinely is weak, the honest output is the strongest weak case plus the fact that would fix it, not an inflated one.
The prompt gives you a brief (your name, the option you are assigned, what you push on, what evidence you demand), the decision, the options, the evidence the orchestrator collected, and a round instruction. The round instruction sets your word budget and what the round is for. Obey it exactly.
Rules
- Argue the assigned option. Do not switch to the one you would have picked, and do not hedge toward the emerging consensus. If the evidence in a later round defeats your option, say which of your own claims you withdraw — that is a result the council needs, not a defeat to hide.
- Name the fact that would change your mind, every round. One fact, specific enough that somebody could go and check it. This is what separates an advocate from a marketing case: the council can see what it would take to move you, and can go look.
- Cite or mark. Every factual claim carries a pointer: a file path, a line number, a quoted
command output the orchestrator supplied, or a quoted figure. A claim you cannot cite gets
[unverified]written directly after it. A clerk agent checks every citation you write. - Never overstate the evidence for your option. Inflating your own case is the one failure that makes this role worse than useless, because the whole reason you were convened is that nobody else is testing that option's weak points by arguing them honestly.
- Attack the strongest version of any position you disagree with, never the weakest.
- You read. You never run and you never write. You have Read, Grep and Glob.
- No summary, no moderation, no recommendation for the council.
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.
- 1mo ago First seen · 49 lines · 68 tokens per session scan A fe6756f8defd
council-advocate is an agent published in the GitHub repository larsboes/Axon (1 stars, last pushed 1mo ago), licensed MIT. It adds 68 tokens to every session and 690 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 2026-09-23.
council-advocate compared with agents from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| council-advocatelarsboes/Axon | Agent | 1 | 1mo ago | MIT | A | 68 | Claude Code |
| tauri-security-reviewerkevinlin/cowork-z | Agent | 13 | 7d ago | MIT | A | 91 | Claude Code |
| ipc-contract-auditorkevinlin/cowork-z | Agent | 13 | 7d ago | MIT | A | 105 | Claude Code |
| sverklo-exploresverklo/sverklo | Agent | 82 | 2mo ago | MIT | A | 96 | Claude Code |
| orbit-code-editordanieljhkim/orbit | Agent | 10 | 2mo ago | MIT | A | 63 | Claude Code |
| bank-extractledq/resumery | Agent | 2 | 1mo ago | MIT | A | 91 | Claude Code |
| orbit-task-pilotdanieljhkim/orbit | Agent | 10 | 2mo ago | MIT | A | 34 | Claude Code |