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 skills add toflow-ai/toflow-mcp --skill toflow-manage-conversationsgit clone --depth 1 https://github.com/toflow-ai/toflow-mcpWrote 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/toflow-ai/toflow-mcp/toflow-manage-conversations)<a href="https://agentmods.dev/skills/toflow-ai/toflow-mcp/toflow-manage-conversations"><img src="https://agentmods.dev/badge/skills/toflow-ai/toflow-mcp/toflow-manage-conversations/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/skills/toflow-ai/toflow-mcp/toflow-manage-conversations"><img src="https://agentmods.dev/badge/skills/toflow-ai/toflow-mcp/toflow-manage-conversations.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 | $0.00052 | $0.00590 |
| Opus 5.5 | $0.00021 | $0.00236 |
| Sonnet 5.5 | $0.00010 | $0.00118 |
| Haiku 4.5 | $0.00005 | $0.00059 |
Grade A, and why
toflow-manage-conversations 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 — 30 lines — stays where its author put it; the contents beside it link to each section on GitHub.
toflow — Manage LinkedIn & WhatsApp Conversations
Handles one-off conversation actions on LinkedIn and WhatsApp. For messages that are part of a multi-step outreach flow, use the toflow-create-sequence skill instead.
Steps
- Call
list_message_accountsto see which LinkedIn and WhatsApp accounts are connected before sending anything. - Call
list_message_threadsto find existing conversations with a contact, or to review recent activity across channels. - Call
get_message_threadto read the full history of a specific thread before replying — context matters more on these channels than email, since messages are short and informal. - Check connection status (via the prospecting tools in
toflow-prospect-and-list) before choosing how to reach someone:- Already connected →
send_linkedin_message. - Not connected →
send_connection_request, optionally with a note (ask the user first — a request without a note generally allows a higher daily send volume than one with a note attached). - Premium/InMail-style reach to someone not in-network and not accepting a connection request →
send_inmail. Don't use this on an existing connection; usesend_linkedin_messageinstead. - WhatsApp →
send_whatsapp_message.
- Already connected →
send_draft_message/delete_message_draftoperate on a previously generated message draft (for example, AI-generated enrollment content awaiting review — seetoflow-create-sequence). Always show the draft's actual content to the user before callingsend_draft_message; never assume what a draft contains.- If the user has multiple connected accounts for a channel, use
set_primary_accountto set the default, or ask which one to use per message if it varies by contact.
Guardrails
- Always show drafted LinkedIn/WhatsApp content for approval before sending — same as email.
- No em dashes, no superlatives ("best-in-class", "game-changing"), no pressure language ("act now") in generated content — these read as AI-written and hurt reply rates on informal channels even more than on email.
- Open with a specific observation (their post, role change, company milestone), not a pitch; ask one easy question rather than a meeting ask in the first message.
- Don't over-message a thread — check
get_message_threadfor recent activity before sending another follow-up. - Don't send InMail to someone already a 1st-degree connection, and don't send a second connection request to someone already connected or already pending — check status first.
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 · 30 lines · 52 tokens per session scan A b23da8ec7c1f
toflow-manage-conversations is a skill published in the GitHub repository toflow-ai/toflow-mcp (6 stars, last pushed 5d ago), licensed MIT. It adds 52 tokens to every session and 590 once invoked, about $0.0002 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 2026-10-10.
Other skills, from other repositories
job-posting-intent
Detect buying intent from job postings. When a company posts a job in your problem area, they've allocated budget and are actively thinking about the problem. This skill finds those companies, qualifies them, extracts personalization context, and outputs everything to a Google Sheet. Does NOT do outreach — just…
ai-cold-email
Generate personalized cold outreach emails using AI.
social-lead-gen
Lead generation from social media — find high-intent buyers in live Twitter, Instagram, and Reddit conversations. Auto-researches your product, generates targeted search queries, and discovers people actively looking for solutions you offer. Social selling and prospecting powered by billions of indexed posts.
end-to-end-leadership-change-signal
Full pipeline from leadership change discovery to outreach-ready campaign. Scans the web for recent executive moves (no company list needed), verifies and enriches via Apollo, evaluates relevance to your product, drafts personalized outreach tuned to the new leader's timing window and background, and packages for your…
end-to-end-news-signal
Full pipeline from news signal discovery to outreach-ready campaign. Scans the web for recent news events — regulation changes, acquisitions, product launches, market shifts, crises, partnerships, earnings, and any other public event — qualifies affected companies against your ICP, identifies the connection between…
end-to-end-funding-signal
Full pipeline from funding signal discovery to outreach-ready campaign. Scans the web for recent funding announcements (no company list needed), qualifies against your company context, finds decision-makers, drafts personalized email sequences, and packages for your outreach tool. Combines funding-signal-monitor and…