Goal
Fetch recent mentions, replies, and comments directed at the account to monitor and respond to community interactions.
Which Agents Use This
- engagement — Monitor for new mentions since last run, identify which need responses
- moderator — Check for complaints or urgent issues requiring human escalation
- custom — Any agent handling community engagement
Hard Rules
- MUST fetch mentions directed at the account (not general brand mentions)
- SHOULD calculate sentiment and priority for triage
- SHOULD flag has_response to avoid duplicate replies
- Priority levels help agents decide what to respond to first
Steps
1. Validate input
- Verify account_id exists and belongs to workspace
- Validate since timestamp if provided
- Apply limit (max 100)
2. Fetch mentions from platform
- Query platform API for mentions/replies/@mentions
- Filter by since timestamp
- Apply filter (all, unanswered, flagged)
3. Analyze mentions
- Determine sentiment (positive, neutral, negative)
- Calculate priority (normal, high, urgent) based on:
- Sentiment (complaints = higher priority)
- Keywords (refund, broken, help = urgent)
- Author influence/follower count
- Check has_response
4. Return mentions array
- Include mention metadata, author, content, timestamps
- Include analysis (sentiment, priority, has_response)
Output
{
"success": true,
"count": 12,
"mentions": [
{
"mention_id": "uuid",
"author": "@username",
"content": "Hey @brand, how do I reset my password?",
"platform": "twitter",
"created_at": "2026-02-20T09:15:00Z",
"sentiment": "neutral",
"has_response": false,
"priority": "high"
}
]
}
Failure Handling
- If account_id not found: return error "Account not found"
- If platform API fails: return cached mentions with warning
- If no mentions found: return empty array with count: 0
Example Usage
Scenario: Engagement agent checking for new mentions
{
"account_id": "uuid",
"since": "2026-02-20T08:00:00Z",
"filter": "unanswered"
}
Result: Receive list of unanswered mentions since 8am for response prioritization
1---2name: bolta-get-mentions3version: 2.0.04description: Fetch recent mentions, replies, and comments directed at the account for community engagement monitoring5category: engagement6roles_allowed: [Viewer, Creator, Editor, Admin]7agent_types: [engagement, custom]8safe_defaults: {}9tools_required:10 - platform_api_integration11inputs_schema:12 type: object13 required: [account_id]14 properties:15 account_id: { type: string, description: "Social account UUID" }16 since: { type: string, description: "ISO timestamp - fetch mentions since this time" }17 limit: { type: number, description: "Max mentions to return", default: 50 }18 filter: { type: string, enum: [all, unanswered, flagged], default: all }19outputs_schema:20 type: object21 properties:22 success: { type: boolean }23 count: { type: number }24 mentions: 25 type: array26 items:27 type: object28 properties:29 mention_id: { type: string }30 author: { type: string }31 content: { type: string }32 platform: { type: string }33 created_at: { type: string }34 sentiment: { type: string }35 has_response: { type: boolean }36 priority: { type: string }37organization: bolta.ai38author: Bolta Team39---4041## Goal42Fetch recent mentions, replies, and comments directed at the account to monitor and respond to community interactions.4344## Which Agents Use This45- **engagement** — Monitor for new mentions since last run, identify which need responses46- **moderator** — Check for complaints or urgent issues requiring human escalation47- **custom** — Any agent handling community engagement4849## Hard Rules501. MUST fetch mentions directed at the account (not general brand mentions)512. SHOULD calculate sentiment and priority for triage523. SHOULD flag has_response to avoid duplicate replies534. Priority levels help agents decide what to respond to first5455## Steps5657### 1. Validate input58- Verify account_id exists and belongs to workspace59- Validate since timestamp if provided60- Apply limit (max 100)6162### 2. Fetch mentions from platform63- Query platform API for mentions/replies/@mentions64- Filter by since timestamp65- Apply filter (all, unanswered, flagged)6667### 3. Analyze mentions68- Determine sentiment (positive, neutral, negative)69- Calculate priority (normal, high, urgent) based on:70 - Sentiment (complaints = higher priority)71 - Keywords (refund, broken, help = urgent)72 - Author influence/follower count73- Check has_response7475### 4. Return mentions array76- Include mention metadata, author, content, timestamps77- Include analysis (sentiment, priority, has_response)7879## Output80```json81{82 "success": true,83 "count": 12,84 "mentions": [85 {86 "mention_id": "uuid",87 "author": "@username",88 "content": "Hey @brand, how do I reset my password?",89 "platform": "twitter",90 "created_at": "2026-02-20T09:15:00Z",91 "sentiment": "neutral",92 "has_response": false,93 "priority": "high"94 }95 ]96}97```9899## Failure Handling100- If account_id not found: return error "Account not found"101- If platform API fails: return cached mentions with warning102- If no mentions found: return empty array with count: 0103104## Example Usage105106### Scenario: Engagement agent checking for new mentions107```json108{109 "account_id": "uuid",110 "since": "2026-02-20T08:00:00Z",111 "filter": "unanswered"112}113```114**Result:** Receive list of unanswered mentions since 8am for response prioritization
boltaai