curator

curator is an agent for Claude Code from southlab-ai/Claude-Plugin-Marketplace. It costs 21 tokens per session (307 once invoked), scanned A, original, MIT.

A memory-curation agent for The Council, a team of cooperating agents that share stored project notes.

In plain words
What is it for?
It is for reviewing role memories, merging related entries, lowering the detail of less important entries, and pruning entries while preserving pinned ones.
Why use it?
It reduces duplicated, outdated, and overly detailed memory so teammates load more relevant context.

Agent for Claude Code

Written for Claude Code: shipped in a Claude Code plugin.

Part of the the-council plugin — 8 skills, 8 agents shipped together

Good fit It is for reviewing role memories, merging related entries, lowering the detail of less important entries, and pruning entries while preserving pinned ones.

Compare 6 agents from other repositories ↓
Install with agentmods
npx agentmods add agents/southlab-ai/claude-plugin-marketplace/curator
Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

Clone the repo
git clone --depth 1 https://github.com/southlab-ai/Claude-Plugin-Marketplace

Made for: Claude Code.

Or install the-council, the plugin that ships this one along with the rest of its 8 skills, 8 agents.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for curator

README.md
[![agentmods](https://agentmods.dev/badge/agents/southlab-ai/claude-plugin-marketplace/curator/github.svg)](https://agentmods.dev/agents/southlab-ai/claude-plugin-marketplace/curator)
Your own site
<a href="https://agentmods.dev/agents/southlab-ai/claude-plugin-marketplace/curator"><img src="https://agentmods.dev/badge/agents/southlab-ai/claude-plugin-marketplace/curator/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for curator

Your own site · 80×15
<a href="https://agentmods.dev/agents/southlab-ai/claude-plugin-marketplace/curator"><img src="https://agentmods.dev/badge/agents/southlab-ai/claude-plugin-marketplace/curator.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 21 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 307 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00021 $0.00307
Opus 5 $0.00010 $0.00153
Sonnet 5 $0.00004 $0.00061
Haiku 4.5 $0.00002 $0.00031

Measured 10d ago against content hash 38e585fd1098, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-10, from the pricing page.

Security

Grade A, and why

curator scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 10d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

plugins/the-council/agents/curator.md · 35 lines

What it actually says

Memory Curator

You compact The Council's memory files so teammates load only relevant, deduplicated context.

Process

For each role (strategist, critic, hub):

  1. Call council_memory_load with project_dir and max_tokens=8000 to see current active entries
  2. Read the role's log file ({role}-log.md) for full history
  3. Identify:
    • Duplicates: same insight in multiple entries -> keep most precise
    • Superseded: overridden by later decisions -> lower importance
    • Mergeable: related insights -> combine into one entry
  4. Build compacted entries array (JSON) with updated importance scores
  5. Call council_memory_compact with role and the JSON array

Rules

  • Target: under 20 entries per role
  • Importance >= 7: keep at detail_level 3 (full text)
  • Importance 4-6: reduce to detail_level 2 (summary)
  • Importance 1-3: reduce to detail_level 1 (headline only)
  • Pinned entries: never prune below detail_level 2
  • NEVER modify archive files (logs, decisions.md, lessons.jsonl)

Output

Report to team lead: entries before -> after for each role, what was removed and why.

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 10d ago First seen · 35 lines · 21 tokens per session scan A 38e585fd1098

Subscribe to this mod's changes

curator is an agent published in the GitHub repository southlab-ai/Claude-Plugin-Marketplace (2 stars, last pushed 1mo ago), licensed MIT. It adds 21 tokens to every session and 307 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Related

Other agents, from other repositories

context-advisor

Use this agent to assess context window health and recommend compaction or session management strategies. Examples: User: "My session feels slow and repetitive" Assistant: "I'll use the context-advisor agent to assess context health." User: "Should I compact or start fresh?" Assistant: "Let me launch the…

Sagargupta16/claude-skills · 76 tokens

starlight-energy-cost

Energy project economics — LCOE, payback, IRR, NPV, financing structures, tariffs, incentives. Translates a sized system into honest financial outcomes across purchase / lease / PPA structures.

frankxai/Starlight-Intelligence-System · 0 tokens

starlight-energy-installer

Installer-side operations — lead intake, site survey, permit/AHJ coordination, work orders, commissioning, customer handoff. The agent that runs the install business, not the install itself.

frankxai/Starlight-Intelligence-System · 0 tokens

starlight-energy-operations

Post-install operations — production monitoring, fault diagnosis, predictive maintenance, warranty claim lifecycle, customer performance reporting.

frankxai/Starlight-Intelligence-System · 0 tokens

starlight-energy-recovery

Disaster / outage / deplatform recovery — grid-down protocols, installer-bankruptcy bridges, monitoring-shutdown migrations, warranty-failure paths, off-grid bridging.

frankxai/Starlight-Intelligence-System · 0 tokens

starlight-energy-sizing

Energy system capacity and component sizing — solar PV, battery storage, EV chargers, heat pumps. Translates load profiles into kW/kWh component specs with derating, capacity factor, and peak demand grounded in real-world physics.

frankxai/Starlight-Intelligence-System · 0 tokens