| name | kabbo |
| description | Manage your academic publication pipeline via the Kabbo MCP server. List, search, create, update, move, and analyse publications; track stalled papers and reminders; export/import BibTeX; link a data repo or Overleaf project; review team progress. |
| allowed-tools | Bash, Read, Write, Edit, Glob, Grep |
| user-invocable | true |
Kabbo — Publication Pipeline Skill
Kabbo (https://kabbo.app) tracks academic papers through seven stages:
Idea → Draft → Submitted → Revise & Resubmit → Resubmitted → Accepted →
Published. This skill drives the Kabbo MCP server.
Connection
If you installed the Kabbo plugin (/plugin install kabbo@econtools), the
MCP server is already configured — you were prompted for your API key at install.
Otherwise add this to ~/.claude/settings.json (replace YOUR_API_KEY):
{
"mcpServers": {
"kabbo": {
"type": "http",
"url": "https://jydnsbaztvmjkebhmoia.supabase.co/functions/v1/mcp-server",
"headers": { "x-api-key": "YOUR_API_KEY" }
}
}
}
The same server backs Codex (via AGENTS.md) and Gemini CLI (in
~/.gemini/settings.json, using httpUrl instead of url). Generate a key at
kabbo.app → Settings → AI Integration → Create Key.
Slash commands (plugin)
/kabbo:status — quick pipeline overview.
Tools
Core CRUD
- list_publications — list with optional
query, stage, limit, offset
- get_publication — one by
id
- create_publication — create (title required; stage defaults to "idea"; upserts by title)
- update_publication — update any field by
id
- move_stage — move to a new
stage by id
- delete_publication — soft-delete (bin) by
id
Analytics & insights
- get_pipeline_summary — counts by stage, stalled (30+ days), recently updated
- get_stalled_papers — papers inactive for N days (default 30)
- get_analytics — velocity, avg time per stage, breakdowns by author/theme/grant/year/output_type
- get_activity_log — recent activity from all sources (api, mcp)
Search & batch
- search_publications — multi-field search (title, authors, notes, themes, grants, year)
- bulk_update — array of
{id, ...updates}
Workflow & metadata
- link_repo — attach a
github_repo URL and/or overleaf_url to a publication (where the data or source lives)
- set_target_journal — set the intended journal
- import_bibtex — create publications from a BibTeX string
- manage_related_papers — list/add/remove related papers
- manage_data_sources — list/add/remove data sources
Reminders, team, export, notes
- manage_reminders —
action: list | create | complete | delete
- get_team_summary — per-member pipeline (needs
team_id)
- export_bibtex — BibTeX for selected/all publications (filter by stage/year/ids)
- add_note — append a timestamped note without overwriting
Resources (read as context)
kabbo://pipeline/summary · kabbo://publications ·
kabbo://publication/{id} · kabbo://activity/recent
Prompts (slash commands in Claude Code)
morning_checkin · weekly_review · annual_report (arg: year) ·
submission_prep (arg: title) · stalled_triage
Updating the pipeline on request
This is the main job. When the user says something like "I just submitted the
climate paper to the AER — update Kabbo":
search_publications (or list_publications) to find the card by title.
update_publication to fill in the metadata — target_journal, authors,
themes, notes — and link_repo to record the GitHub/Overleaf URL where
the data or source lives.
move_stage to the correct column (here, submitted).
There is no GitHub/Overleaf automation — you set these fields directly. The repo
and Overleaf links are plain metadata on the card.
Common workflows
- Morning check-in →
get_pipeline_summary
- Weekly review →
get_activity_log (days: 7) + get_pipeline_summary
- Find stalled work →
get_stalled_papers
- Annual review →
get_analytics + export_bibtex
- Import a CV/BibTeX →
import_bibtex
- Batch tag →
list_publications then bulk_update
- Set a reminder →
manage_reminders (action: create, due_date ISO)
- Team oversight →
get_team_summary (team_id required)
Tips
- Search by title before creating —
create_publication upserts by title.
- Stage names use underscores:
revise_resubmit. Aliases accepted: r&r, wip, in-review, forthcoming.
get_pipeline_summary is the best single-call overview.
- For team operations you need the team UUID (from the Kabbo web UI).