gabrielmoreira Skill Path
Before running helper scripts or opening bundled references, set:
export MINUTES_SKILLS_ROOT="$(git rev-parse --show-toplevel)/.agents/skills/minutes"
export MINUTES_SKILL_ROOT="$MINUTES_SKILLS_ROOT/minutes-mcp-recall"
/minutes-mcp-recall
Route a meeting-memory question to the smallest useful Minutes MCP tool. This
skill is tool-name-oriented for MCP hosts such as OpenClaw, Hermes-agent, and
other MCP clients. Use minutes-search instead when the host only has CLI or
shell access.
Read the complete tool map only when the decision table does not cover the request.
Decision table
| User is asking... | Call |
|---|---|
| "What did we say about X?" or "Find that meeting about Y." | search_meetings |
| "What is the full transcript of this specific meeting?" | get_meeting after finding its id or date with list_meetings or search_meetings |
| About what is happening right now, mid-call | read_live_transcript with a cursor or time window |
| "What do I still owe?" or "Who owes me?" | track_commitments |
| "What is my history with this person?" | get_person_profile or relationship_map |
| "Have I contradicted myself?" or "Which decisions are stale?" | consistency_report |
| Broad research across many meetings | research_topic |
Tool-name prefix by host
- OpenClaw and generic MCP clients use bare tool names such as
search_meetings. - Hermes-agent prefixes each tool with
mcp_minutes_, sosearch_meetingsbecomesmcp_minutes_search_meetings.
Apply the same prefix rule to every tool in the decision table and tool map.
Anti-patterns
- Do not call
get_meetingspeculatively to browse. It returns a full transcript. Usesearch_meetingsorlist_meetingsfirst to find the right id or date and preserve context budget. - Do not call
read_live_transcriptwithout a cursor after one is available. Replaying the whole transcript wastes tokens and duplicates content already in context. - Do not use capture, dictation, processing, annotation, or speaker-mutation tools for a recall question.