intercom-automation

Automate Intercom tasks via Rube MCP (Composio): conversations, contacts, companies, segments, admins. Always search tools first for current schemas.

Install with Tessl CLI

npx tessl i github:davepoon/buildwithclaude --skill intercom-automation

What are skills?

1.55x

Quality

52%

Does it follow best practices?

Impact

92%

1.55x

Average score across 3 eval scenarios

Optimize this skill with Tessl

npx tessl skill review --optimize ./plugins/all-skills/skills/intercom-automation/SKILL.md

SKILL.md

Review

Evals

Evaluation results

100%

35%

Conversation Triage Automation Script

Conversation management workflow with admin ID resolution

Criteria

Without context

With context

RUBE_SEARCH_TOOLS first

100%

RUBE_MANAGE_CONNECTIONS setup

100%

Admin ID resolution first

75%

100%

Structured conversation search

100%

Correct conversation search fields

100%

AND operator for multiple filters

100%

Correct assignment parameters

100%

Admin IDs from LIST_ALL_ADMINS result

100%

Teams distinguished from admins

50%

100%

Pagination handling

100%

Action logging

100%

Without context: $1.2351 · 4m 39s · 28 turns · 134 in / 8,881 out tokens

With context: $0.5738 · 2m 25s · 23 turns · 28 in / 8,299 out tokens

100%

61%

Lead Qualification Audit Script

Structured contact search with pagination and external ID lookup

Criteria

Without context

With context

RUBE_SEARCH_TOOLS first

100%

Connection verification

100%

Structured search query

100%

AND operator for multi-filter

100%

Correct contact filter fields

62%

100%

Correct operators used

25%

100%

Cursor-based pagination

83%

100%

pagination.starting_after parameter

100%

External ID exact lookup

60%

100%

Case-sensitive external IDs

100%

Role=lead filter value

100%

Without context: $0.7008 · 3m 10s · 41 turns · 120 in / 10,798 out tokens

With context: $0.5590 · 2m 18s · 20 turns · 25 in / 8,864 out tokens

78%

Critical Issue Incident Response Automation

Idempotent conversation workflow with internal notes and retry logic

Criteria

Without context

With context

RUBE_SEARCH_TOOLS first

Connection verified active

28%

Admin lookup before operations

70%

Internal note uses message_type note

100%

Customer reply uses message_type comment

100%

admin_id in reply/note/close

100%

70%

Exponential backoff on 429

100%

Idempotency mechanism

100%

Idempotency explained in design_notes

100%

Correct close parameters

100%

62%

HTML content note

100%

Without context: $0.5680 · 2m 29s · 30 turns · 36 in / 8,361 out tokens

With context: $0.4518 · 1m 59s · 18 turns · 426 in / 6,759 out tokens

Evaluated: 4 days ago
Agent: Claude Code
Model: Claude Sonnet 4.6

Table of Contents

Conversation Triage Automation Script Lead Qualification Audit Script Critical Issue Incident Response Automation

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.