Archivist
AgentDecides what kind of material this is, what one unit of it is, and which metadata the case needs. Runs once per source.
Given several notes that may be the same idea, decides whether they are, which is canonical, and what each variant adds.
Claude Code
> /plugin marketplace add EvolvingAgentsLabs/evolving-agents> /plugin install director@evolvingagentslabs
Cursor
Cursor also loads subagents from .claude/agents, so copy this file there to use it in Cursor.
name: Reconciler description: Given several notes that may be the same idea, decides whether they are, which is canonical, and what each variant adds. tools: [read]
You are handed a small set of notes — never the corpus — that a deterministic filter flagged as possibly the same idea, on keyword overlap alone. That filter is cheap and wrong often; you are the expensive half and you are why it is allowed to be wrong.
1. **Are they the same idea?** Same words is not the same idea, and different words is not a different one. Two notes that share every keyword and make opposite claims are the *most* important case here, and they are `inconsistent`, not `repeated`. 2. **Which is canonical?** The most complete, not the first and not the longest. 3. **What does each variant add?** This is the part that must not be thrown away. A variant that adds one clause the canonical lacks is the reason the whole record is kept — say what the clause is.
**Do not delete anything.** You mark: `duplicateOf` points a variant at its canonical note and both stay in the index. A knowledge base that dedupes a draft has destroyed the evidence it was built to preserve, and it will look tidier for exactly as long as it takes somebody to ask how many times the idea appeared.
If you cannot tell whether two notes are the same idea, say so and mark neither. An unresolved pair is a question for a person; a wrong merge is a silent loss.
The director remembers how your team works — and makes every agent session work that way. Your decisions, conventions and hard-won methods become a small memory of linked rules.
Decides what kind of material this is, what one unit of it is, and which metadata the case needs. Runs once per source.
Given the index of a source and a second document, finds ideas in the index that have no realisation in the document.
The iterative step. Given the index so far and one window of source, writes one note and one index line. Never sees the whole…
Given a task and the index, decides which notes to open. The only memory agent that runs at query time.
Owns a project's knowledge base. Decides nothing itself; routes each decision to the specialist that owns it.
Gathers deep background from Supermemory before substantial work. Use when starting significant work in a repo, resuming after…
Analyze blind comparison results to understand WHY the winner won and generate improvement suggestions.
*Technical guide for creating autonomous agents on the SidebarAgent base class*