diff --git a/productivity/.mcp.json b/productivity/.mcp.json
index 0097bab3..8e404cd5 100644
--- a/productivity/.mcp.json
+++ b/productivity/.mcp.json
@@ -43,6 +43,14 @@
     "gmail": {
       "type": "http",
       "url": ""
+    },
+    "fireflies": {
+      "type": "http",
+      "url": "https://api.fireflies.ai/mcp"
+    },
+    "grain": {
+      "type": "http",
+      "url": "https://api.grain.com/_/mcp"
     }
   }
 }
diff --git a/productivity/CONNECTORS.md b/productivity/CONNECTORS.md
index 537f5093..f2185630 100644
--- a/productivity/CONNECTORS.md
+++ b/productivity/CONNECTORS.md
@@ -13,6 +13,7 @@ Plugins are **tool-agnostic** — they describe workflows in terms of categories
 | Chat | `~~chat` | Slack | Microsoft Teams, Discord |
 | Email | `~~email` | Microsoft 365 | — |
 | Calendar | `~~calendar` | Microsoft 365 | — |
+| Meeting transcription | `~~conversation intelligence` | Fireflies, Grain | Gong, Granola, Chorus, Otter.ai |
 | Knowledge base | `~~knowledge base` | Notion | Confluence, Guru, Coda |
 | Project tracker | `~~project tracker` | Asana, Linear, Atlassian (Jira/Confluence), monday.com, ClickUp | Shortcut, Basecamp, Wrike |
 | Office suite | `~~office suite` | Microsoft 365 | — |
diff --git a/productivity/skills/start/SKILL.md b/productivity/skills/start/SKILL.md
index 36be8ceb..374beed5 100644
--- a/productivity/skills/start/SKILL.md
+++ b/productivity/skills/start/SKILL.md
@@ -97,6 +97,7 @@ Gather data from available MCP sources:
 - **Email:** Sent messages, recipients
 - **Documents:** Recent docs, collaborators
 - **Calendar:** Meetings, attendees
+- **Meeting transcripts** (if `~~conversation intelligence` MCP available): Retrieve the user's most recent ~20 attended meetings — pull attendee names, recurring meeting series, project codenames mentioned in transcripts, and AI-generated meeting notes. Meeting transcripts are the highest-density source of workplace shorthand (nicknames, project codenames, internal terms) — favor them when available.
 
 Build a braindump of people, projects, and terms found. Present findings grouped by confidence:
 - **Ready to add** (high confidence) — offer to add directly
diff --git a/productivity/skills/update/SKILL.md b/productivity/skills/update/SKILL.md
index ee86a1f3..217f3b46 100644
--- a/productivity/skills/update/SKILL.md
+++ b/productivity/skills/update/SKILL.md
@@ -30,6 +30,7 @@ Read `TASKS.md` and `memory/` directory. If they don't exist, suggest `/producti
 
 Check for available task sources:
 - **Project tracker** (e.g. Asana, Linear, Jira) (if MCP available)
+- **Conversation intelligence** (e.g. Fireflies, Grain, Gong) (if `~~conversation intelligence` MCP available) — pull open action items assigned to the user from recent meetings; surface under a "Meeting Action Items" section in TASKS.md
 - **GitHub Issues** (if in a repo): `gh issue list --assignee=@me`
 
 If no sources are available, skip to Step 3.
@@ -91,6 +92,7 @@ Tasks often contain richer context than memory. Extract and update:
 - **Status changes** ("launch done") → update project status, demote from CLAUDE.md
 - **Relationships** ("Todd's sign-off on Maya's proposal") → cross-reference people
 - **Deadlines** → add to project files
+- **Meeting decisions and attendees** (if `~~conversation intelligence` MCP available) → update `memory/` entries for people, projects, and companies mentioned in recent meetings; capture decisions surfaced in meeting notes that shift project status
 
 ### 7. Report
 
@@ -112,6 +114,7 @@ Gather data from available MCP sources:
 - **Email:** Search sent messages
 - **Documents:** List recently touched docs
 - **Calendar:** List recent + upcoming events
+- **Meeting transcripts** (if `~~conversation intelligence` MCP available): List recent attended meetings, retrieve transcripts and AI-generated notes, surface action items and decisions
 
 ### Extra Step: Flag Missed Todos