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/evoclaw/amplify/multi-round-deliberationnpx skills add EvoClaw/amplify --skill multi-round-deliberationgit clone --depth 1 https://github.com/EvoClaw/amplifyWrote 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/evoclaw/amplify/multi-round-deliberation)<a href="https://agentmods.dev/skills/evoclaw/amplify/multi-round-deliberation"><img src="https://agentmods.dev/badge/skills/evoclaw/amplify/multi-round-deliberation.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 5.1 | $0.00047 | $0.02847 |
| Opus 5 | $0.00023 | $0.01424 |
| Sonnet 5 | $0.00009 | $0.00569 |
| Haiku 4.5 | $0.00005 | $0.00285 |
Grade A, and why
multi-round-deliberation 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 6d 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 — 299 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Multi-Round Deliberation Protocol
Overview
A single round of feedback is NOT a discussion. Real deliberation requires iteration: agents identify issues → someone fixes them → agents verify the fix → repeat until resolved. This protocol standardizes this loop across all phases.
The Problem with Single-Round Feedback
WRONG (current anti-pattern):
3 agents give opinions → main agent "synthesizes" → done
→ Issues may be noted but never actually resolved
→ No verification that fixes addressed the concerns
RIGHT (this protocol):
3 agents give opinions → main agent modifies artifact →
agents re-check → if unresolved, modify again →
repeat until convergence or max rounds
Core Protocol
Shared Value Framework (prevents divergence)
All agents in a deliberation share these values, injected into every prompt:
SHARED VALUES (inject into every agent prompt in this deliberation):
═══════════════════════════════════════════════════════════════════
Target venue: [from research-anchor.yaml]
Value proposition: [from research-anchor.yaml]
Research type: [from research-anchor.yaml]
OPTIMIZATION TARGET: "Would this survive peer review at [target venue]?"
SCORING RUBRIC (default — phases may override; see phase-specific sections):
- PASS: No fatal or major issues. Ready to proceed.
- CONDITIONAL: Major issues exist but are addressable. Needs another round.
- FAIL: Fatal issues. Fundamental rethinking needed.
CONVERGENCE RULE: Deliberation ends when ALL agents score PASS,
or when max rounds are reached and remaining issues are presented to user.
ANTI-DIVERGENCE: If you disagree with another agent's feedback,
state specifically WHY and propose a CONCRETE alternative.
"I disagree" without a counter-proposal is not constructive.
═══════════════════════════════════════════════════════════════════
Deliberation Loop
┌─────────────────────────────────────────────────────┐
│ ROUND N │
│ │
│ 1. ASSESS (parallel) │
│ Dispatch ALL agents with current artifact │
│ Each returns: issues + verdict (PASS/COND/FAIL) │
│ │
│ 2. CHECK CONVERGENCE │
│ All PASS? → END (consensus reached) │
│ Any FAIL or COND + round < max? → go to step 3 │
│ Max rounds (5) reached? → END (present to user) │
│ │
│ 3. MODIFY │
│ Main agent (you) incorporates feedback: │
│ - Fix all issues marked as addressable │
│ - For disagreements: choose the stronger argument│
│ - Update the artifact │
│ │
│ 4. FULL RE-ASSESS (next round) │
│ Dispatch ALL agents again with: │
│ - The complete modified artifact │
│ - Summary of changes made since last round │
│ - Previous round's issues for reference │
│ All agents review the full artifact — not just │
│ their own previous concerns — because │
│ modifications can introduce new issues. │
│ │
│ → Back to step 2 │
└─────────────────────────────────────────────────────┘
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.
- 6d ago First seen · 299 lines · 47 tokens per session scan A f9900cb1351e
multi-round-deliberation is a skill published in the GitHub repository EvoClaw/amplify (12 stars, last pushed 6mo ago), licensed MIT. It adds 47 tokens to every session and 2,847 once invoked, about $0.0002 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 2026-08-30.
Other skills, from other repositories
dingtalk_channel_connect
Use a headed browser to automatically complete DingTalk channel integration for QwenPaw. Applicable when the user mentions DingTalk, developer console, Client ID, Client Secret, bot, Stream mode, binding or configuring a channel. Supports pausing when a login page is detected and resuming after the user logs in.
当用户需要对PDF文件进行任何操作时，请使用此技能。包括从 PDF 中读取或提取文本/表格、合并多个 PDF、拆分 PDF、旋转页面、添加水印、创建新PDF、填写PDF表单、加密/解密 PDF、提取图片，以及对扫描版 PDF 进行 OCR 使其可搜索。如果用户提到 .pdf 文件或要求生成 PDF，请使用此技能。.
make_plan
For external plan request scenarios, guides the Agent to request a clear, actionable, step-by-step plan from a stronger Agent via listagents and chatwithagent, emphasizing that the plan is executed by the requester, not by the consulted Agent.
new
Create a new project to start development quickly.
oss-upload
Upload local files to Tencent COS (oss.1024code.com CDN) using coscli. Use when user wants to upload a file to CDN/OSS, or deploy static assets.
officecli-word-form
Use this skill to create fillable Word forms (.docx) with real Content Controls (SDT) + legacy FormField checkboxes + MERGEFIELD mail-merge placeholders + document protection. Trigger on: 'fillable form', 'form fields', 'content controls', 'SDT', 'word form', 'fill in', 'only editable fields', 'protect document'…