agentmods.dev

seance

seance is a command for coding agents from JokeM543/seance. It costs 24 tokens per session (658 once invoked), scanned A, original, MIT.

A command that hands the current task to a new Claude chat with a written summary. The new chat starts without the original conversation, so the summary carries the necessary context forward.

In plain words
What is it for?
It is for summarizing the current work, saving the summary in the project’s .claude folder, and opening a new chat prepared to continue it.
Why use it?
It prevents a long or interrupted session from losing the task details when work moves to a fresh chat.

Command

Installs and runs on its own, but its text points at files inside its plugin — anything it tells you to read at a ${CLAUDE_PLUGIN_ROOT} path is only there once the plugin is installed. Installing the plugin gets both.

Part of the seance plugin — 1 command, 1 hook shipped together

not rated 5 2mo ago A scan 24 tokens original MIT
View the source file JokeM543/seance
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.

agentmods
npx agentmods add commands/jokem543/seance/seance
Clone the repo
git clone --depth 1 https://github.com/JokeM543/seance

Or install seance, the plugin that ships this one along with the rest of its 1 command, 1 hook.

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 seance

README.md
[![agentmods](https://agentmods.dev/badge/commands/jokem543/seance/seance.svg)](https://agentmods.dev/commands/jokem543/seance/seance)
Your own site
<a href="https://agentmods.dev/commands/jokem543/seance/seance"><img src="https://agentmods.dev/badge/commands/jokem543/seance/seance.svg" alt="Measured on agentmods" height="20"></a>
Per session 24 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 658 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
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 $0.00024 $0.00658
Opus 5 $0.00012 $0.00329
Sonnet 5 $0.00005 $0.00132
Haiku 4.5 $0.00002 $0.00066

Measured 4d ago against content hash 1b045d823686, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

seance 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 4d 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.

commands/seance.md · 71 lines

How it starts

The opening of the file, as written. The whole thing — 71 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are conducting a séance: handing this session off to a brand-new Claude chat that starts with no memory of this conversation. The note you write is the only thing it will see, so make it complete enough that the next session can resume the task immediately.

Step 1 — Generate the note and open the new chat

By default, offload the summarizing to a cheaper model so this (expensive) session does not spend a turn composing the note. Run this single command:

node "${CLAUDE_PLUGIN_ROOT}/scripts/seance-summarize.mjs" --open --note "$ARGUMENTS"

It reads this session's transcript, trims it to a small digest, writes ${CLAUDE_PROJECT_DIR}/.claude/seance.md using a separate model (Sonnet by default; configurable via the summarizer plugin option), and opens a fresh Claude chat pre-loaded with it. If it exits 0, you're done — go to Step 2.

Fallback — compose it yourself

If that command fails (non-zero exit) — e.g. claude -p is unavailable, or the user set summarizer = self (the script bails so this path runs) — then compose the note from THIS conversation yourself and open the chat manually:

  1. Write ${CLAUDE_PROJECT_DIR}/.claude/seance.md (create .claude/ if needed) using exactly these sections:

    # Séance note

## Goal
<objective, 1–3 sentences>

## Current state
<what's done; what works / doesn't right now>

## Key decisions & constraints
<choices made, what NOT to redo, conventions, gotchas>

## Files touched
<bullet list of paths, each a one-line note>

## Next steps
<ordered, actionable to-do — start here>

## How to run / test
<exact commands>

    Keep it lean (this is the point): ≤ ~40 lines, bullets only, reference files by path, never paste code or long output, omit empty sections, capture decisions/state — not a transcript. If the user passed an extra note as $ARGUMENTS, fold it into Next steps.

  2. Open the new chat:

    node "${CLAUDE_PLUGIN_ROOT}/scripts/seance-open.mjs" --file "${CLAUDE_PROJECT_DIR}/.claude/seance.md"

Read the full file on GitHub · 71 lines

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. 4d ago First seen · 71 lines · 24 tokens per session scan A 1b045d823686

Subscribe to this mod's changes

seance is a command published in the GitHub repository JokeM543/seance (5 stars, last pushed 2mo ago), licensed MIT. It adds 24 tokens to every session and 658 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.

Building agentsLLManthropicclaudeclaude-codecontext-managementdeveloper-tools
Related

Other commands, from other repositories

decompose

Decompose an RLM execution chain into subtasks and begin processing.

joseairosa/recall · 16 tokens

rlm-status

Check the status and progress of an RLM execution chain.

joseairosa/recall · 15 tokens

load-context

Load a large file or content into the RLM system for chunk-based analysis.

joseairosa/recall · 18 tokens

recall-loop

Periodic memory digest — designed for use with Claude Code's /loop scheduler. Shows recent memories and pending to-dos in a compact, scannable format.

joseairosa/recall · 36 tokens

recall-set-key

Update your Recall API key across every location Claude Code reads it from. Run this whenever you generate a new key from https://recallmcp.com/dashboard/keys.

joseairosa/recall · 0 tokens

recall-update

Manually trigger a Recall plugin update, or toggle auto-update on/off.

joseairosa/recall · 0 tokens