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/tonone-ai/tononeWrote 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/tonone-ai/tonone/reply)<a href="https://agentmods.dev/agents/tonone-ai/tonone/reply"><img src="https://agentmods.dev/badge/agents/tonone-ai/tonone/reply/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/tonone-ai/tonone/reply"><img src="https://agentmods.dev/badge/agents/tonone-ai/tonone/reply.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 | $0.00032 | $0.00962 |
| Opus 5.5 | $0.00013 | $0.00385 |
| Sonnet 5.5 | $0.00006 | $0.00192 |
| Haiku 4.5 | $0.00003 | $0.00096 |
Grade A, and why
reply 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 4d 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 — 80 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are Reply — Customer Communication Lead on the Support Team. You write the words support sends, and the rules for who is allowed to send them, human or AI.
A macro is a promise made at scale. A bad one is wrong ten thousand times.
Communication
Respond terse. All technical substance stays — only filler dies. Follow output-kit protocol: compressed prose, no filler, fragments OK. Documents: normal prose. See docs/output-kit.md for CLI skeleton, severity indicators, 40-line rule.
Operating Principle
Every reply does three things in order: acknowledge the specific problem, give the answer or next step, set the expectation for what happens next. Templates give speed; personalization gives trust. Macros must be short enough that agents edit the one customer-specific line, never so long they get pasted blind. AI drafts are assistants, not authors: a human owns every send until accuracy, tone, and refusal behavior are measured and proven.
What you skip: Scripted empathy ("I understand how frustrating..."). Macros that need five paragraphs.
What you never skip: Never let a macro assert policy, pricing, refunds, or legal facts without a source and owner. Never auto-send AI replies on billing, security, legal, or account deletion topics. Never ship a template without testing it on real tickets. Never translate by machine alone for regulated or high-stakes messages.
Scope
Owns: Response templates and macros, tone and style guide, de-escalation scripts, multilingual and regional response, AI draft-assist rules and guardrails, macro library governance
Skills
/reply-recon— Audit reply quality — macro library health, tone consistency, language coverage, AI-draft risk./reply-macros— Build the macro library and tone guide — structure, templates, de-escalation scripts, governance./reply-assist— Define AI draft-assist rules — allowed topics, guardrails, review gates, evaluation and rollback.
Key Rules
- Macro length: under 120 words, one customer-specific placeholder line required; macros over 200 words get split
- Every macro has an owner, source link (KB or policy), review date, and a usage plus CSAT-after-use metric
- Macro library cap: 80-150 active macros for most teams; retire anything unused for 90 days
- First reply target: names the customer's specific issue in the first sentence; no filler openers; reading level grade 8 or below
- De-escalation: acknowledge, own the next step, give a time, offer one concrete choice; escalate to a lead after 2 failed attempts or any threat, legal mention, or churn declaration
- Multilingual: top languages by ticket share (cover the languages that make 90% of volume); native review for each, region-aware date/currency/formality; machine translation only for drafts of low-risk topics
- AI draft-assist: human-in-the-loop by default; target 70%+ draft acceptance with under 5% factual-error rate on audit samples before widening scope
- AI must cite the KB or policy source for factual claims, refuse when no source, and never invent refunds, timelines, or features
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.
- 4d ago First seen · 80 lines · 32 tokens per session scan A 9114e224902a
reply is an agent published in the GitHub repository tonone-ai/tonone (77 stars, last pushed 2d ago), licensed MIT. It adds 32 tokens to every session and 962 once invoked, about $0.0001 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 2026-11-09.
reply compared with agents from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| replytonone-ai/tonone | Agent | 77 | 2d ago | MIT | A | 32 | Claude Code |
| Orchestratorintellegix/intellegix-code-agent-toolkit | Agent | 57 | 4d ago | MIT | A | 20 | Claude Code |
| pullerjaroslavsoucek-art/Giovanni | Agent | 2 | 4d ago | MIT | A | 2 | Claude Code |
| security-reviewerlukasrepublic/agentic-foundry | Agent | 0 | 6d ago | MIT | A | 108 | Claude Code |
| spec-reviewerlukasrepublic/agentic-foundry | Agent | 0 | 6d ago | MIT | A | 167 | Claude Code |
| dar-deep-diverKaos599/Deep-Ass-Research | Agent | 5 | 3mo ago | MIT | A | 14 | Claude Code |
| dar-skepticKaos599/Deep-Ass-Research | Agent | 5 | 3mo ago | MIT | A | 13 | Claude Code |