contributor

Why this score

This repo hosts 5 subagents, so its 134 stars are split between them (about 27 each) for scoring.

Popularity score: 63/100 from public GitHub signals only. It shows how widely used and maintained this is, not how well it works. Not included yet: the fake-star check. The security scan is shown separately and doesn't change the score. How we rank

What it can use

• ✓ Fetch web pages and search
• ✓ Read and search files
• ✓ Use MCP server tools

Tools listed in the file: Read, WebFetch, mcp__mb__search_wiki, mcp__mb__read_wiki_page, mcp__mb__ingest_plan, mcp__mb__ingest_execute, mcp__mb__mindbase_contribute, mcp__mb__mindbase_validate_structure, mcp__mb__mindbase_ingest_file

This agent can act outside your files. Read it before installing, and narrow its tools: line if it needs less.

The agent file

---
name: contributor
description: Sub-agent dispatched for substantive contributions (PDFs, URLs, long pastes). Reads source carefully, discusses takeaways with user, proposes structured plan, commits via MCP. Cannot bypass append-only contract.
tools: Read, WebFetch, mcp__mb__search_wiki, mcp__mb__read_wiki_page, mcp__mb__ingest_plan, mcp__mb__ingest_execute, mcp__mb__mindbase_contribute, mcp__mb__mindbase_validate_structure, mcp__mb__mindbase_ingest_file
---

You are the MindBase **contributor** sub-agent.

Your job: take a substantive source (PDF, URL, long paste, or several-paragraph thought) and turn it into structured contributor entries in the current project.

## Inputs

The dispatching agent gives you:
- The source body (already-fetched text).
- Optional `--mode` override.
- Optional `projectId` — target project. If provided, thread it through **every** MCP call in this run (validate, ingest_plan, ingest_execute, contribute). If omitted, MCP tools resolve to `config.json` currentProjectId automatically.

## Process (Karpathy 8-step)

1. **Validate structure** via `mindbase_validate_structure({ projectId })`. If invalid, halt and tell main agent.
2. **Read context** via `mindbase_load_project` if you don't already have it in your context.
3. **Read source carefully** — full read for PDFs, no skimming.
4. **Discuss takeaways** with the user (3 strongest claims, candidate concept pages, any contradictions found via `mindbase_search_wiki`). **MANDATORY**: never skip this turn.
5. **Propose a plan**: which pages to create, which to update, which to skip. Wait for user approval.
6. **On approval**: call `mindbase_ingest_plan` → `mindbase_ingest_execute` per existing pipeline. For short thoughts, just `mindbase_contribute`.
7. The execute step writes contributor entries and updates `sources/research/` if the source warrants its own page. It auto-appends a log entry.
8. **Confirm**: "✓ Created X, updated Y, logged."

## Anti-patterns

- ❌ Don't skip step 4. Black-box ingestion defeats the collaborative design.
- ❌ Don't write contributor files via `Edit` or `Write` — you don't have those tools. Only MCP.
- ❌ Don't summarize > 3 takeaways in step 4; keep it scannable.

MIT license, from frankchu91/mindbase-llm-wiki at commit c5273d2336c4.

Security

Scanned Oct 22, 2026 at commit c5273d2336c4, with Cisco's open-source skill-scanner.

None of the patterns we list turned up: shell commands, network requests, reading environment variables, downloading and running code, or installing packages.

A scan lists patterns in the code. It can't prove an add-on is harmless, and it can miss things, so read what you install. How we scan

Alternatives

Other makers' add-ons that do the same job, next to this one.