agentleFS
Sign inStart free
matthewdonsemail-labmatthewdonsemail-lab/log/skills/conversations/SKILL.md

conversations

Skill63 starsChanged yesterdayMITView on GitHub

Store and manage buyers' Marketplace conversations in Twenty - agencyConversations tied to their agencyListing and socialAccount, agencyMessages tied to their conversation. Use when the owner asks about buyers, leads, messages, inbox, replies, "who wrote about this listing", or to record a scanned thread.

The file

Create a free account to mount
Files
SKILL.md

conversations

A buyer who writes about a listing is an agencyConversation on that listing and account. Each message is an agencyMessage in it. The shape, and where each field came from, is in docs/conversations.md.

Hard boundaries

  1. Twenty only. These commands never open a browser and never write to Facebook (rule 8). Replying waits for the owner.
  2. Additive. A field is only ever added through scripts/twenty/twentySetup.mjs, then the client is regenerated (bun run twenty:client). blaster's SMS fields are never renamed or retyped.
  3. Never guess the listing. A thread is tied to a listing by the Facebook listing id, or by a title only one listing on the account has. Otherwise it stays unmatched until log facebook conversation link.
  4. A person's call stands. BOOKED, LOST and SPAM are never moved by a scan.

The routine

log facebook conversations --account <name>                  # newest first
log facebook conversation show <id>                          # every message; * marks unread
log facebook conversation record --account <name> --file threads.json
log facebook conversation read <id>
log facebook conversation lead <id> QUALIFIED
log facebook conversation link <id> --listing <agencyListingId>

record is idempotent: a second run on the same file adds nothing.

Reading the inbox (next)

The scan that writes threads.json from Facebook isn't built (#41). Record the Marketplace inbox with log inspect first (rule 29: goto the inbox, snapshot, open a thread, snapshot), read its HAR with rea for the requests that carry thread and message ids (rule 25), then build the scan from that recording.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free