session-orchestrator

session-orchestrator is a skill for Claude Code, Codex from rocketmandrey/claude-recall. It costs 196 tokens per session (3,960 once invoked), scanned B, original, MIT.

A command-line workflow for finding and continuing earlier Claude Code work sessions. It searches saved session summaries and can either continue the work here, provide a resume command, or open it in another terminal window.

In plain words
What is it for?
It helps locate past work by project, task, tags, or named entities, inspect its context and files, and decide how to resume it. The provided material is partly in Russian, so some details may require interpretation.
Why use it?
It removes the need to remember which session contained a task or to manually search old session files. It also helps avoid creating a duplicate session when one is already open.

Skill for Claude CodeCodex

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 skills/rocketmandrey/claude-recall/skill
Any agent
npx skills add rocketmandrey/claude-recall --skill skill
Clone the repo
git clone --depth 1 https://github.com/rocketmandrey/claude-recall

Made for: Claude Code, Codex.

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 session-orchestrator

README.md
[![agentmods](https://agentmods.dev/badge/skills/rocketmandrey/claude-recall/skill.svg)](https://agentmods.dev/skills/rocketmandrey/claude-recall/skill)
Your own site
<a href="https://agentmods.dev/skills/rocketmandrey/claude-recall/skill"><img src="https://agentmods.dev/badge/skills/rocketmandrey/claude-recall/skill.svg" alt="Measured on agentmods" height="20"></a>
Per session 196 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 3,960 The whole file, excluding the scripts and references it only reads on demand.
Security scan B 1 finding. 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 $0.00196 $0.03960
Opus 5 $0.00098 $0.01980
Sonnet 5 $0.00039 $0.00792
Haiku 4.5 $0.00020 $0.00396

Measured 5d ago against content hash 8d8feefa3459, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade B, and why

session-orchestrator scanned grade B with 1 finding 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 5d 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.

Reads agent configuration directoriesmediumAgent snooping

.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.

`cat ~/.claude/session-recaps/handoffs/<полный-id>.md`.
skill/SKILL.md · 220 lines

How it starts

The opening of the file, as written. The whole thing — 220 lines — stays where its author put it; the contents beside it link to each section on GitHub.

session-orchestrator — найти сессию, погрузиться, продолжить

CLI: recall (см. recall help). Данные: ~/.claude/session-recaps/ — карточка на сессию + единый INDEX.md (строка на сессию). Это первичная точка касания; запасной полнотекст — recall grep.

Шаг 1 — Поиск по индексу (ВСЕГДА сначала здесь)

recall find <основа1> <основа2> ...
  • Ищи по ОСНОВЕ слова, не по словоформе: «Ивановой» → иванов, «выгрузку» → выгрузк.
  • Несколько вариантов сразу: recall find иванов выгрузк постоянн. Выдача уже отранжирована: совпадение в задаче > в #тегах/{сущностях} > в проекте, свежее выше.
  • Строка индекса: <id8> <дата>T<чч:мм> <проект> [status] <задача> #теги {сущности}.
  • [new] = сессия ещё без LLM-рекапа (свежая или умершая грязно; строка = её первый запрос). Это НОРМАЛЬНЫЙ кандидат — часто именно та самая последняя сессия.
  • ⚡LIVE в конце строки = сессия ОТКРЫТА ПРЯМО СЕЙЧАС. НЕ resume и НЕ recall open (раздвоит контекст) — работай адресно: recall psrecall peek/send <tty>.
  • (meta) / #meta = сессия-искатель (сама искала другие сессии) — почти никогда не то, что нужно; ранжируется вниз автоматически.
  • Можно искать прямо по тегу: recall find #rosperevozki (словарь: recall tags).
  • Текущую сессию игнорируй. Если индекс молчит — fallback: recall grep <основы...> (полнотекст по сырым транскриптам; строка: <id8> <дата> <проект> … сниппет …), затем шаг 2 по найденному id.
  • При «подними последнюю/вчерашнюю сессию» сверяй ВРЕМЯ в строке: одинаковых сессий одного проекта много, нужна самая свежая по <дата>T<чч:мм>, не первая попавшаяся.

Несколько кандидатов → селектор

Если уверенного победителя нет (2+ правдоподобных строки) — НЕ выбирай сам. Построй селектор через AskUserQuestion: один вопрос «Какая сессия?», вариант на кандидата (максимум 4 самых свежих/релевантных):

  • label — папка проекта + дата (my-crm · 11.06);
  • description — контекст из строки индекса: задача, статус, ключевые сущности;
  • последним вариантом всегда — «Копнуть глубже»: по выбору открой карточки (recall show) всех кандидатов, при нужде recall grep или сырые transcripts (~/.claude/projects/), и покажи расширенное сравнение.

Read the full file on GitHub · 220 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. 5d ago First seen · 220 lines · 196 tokens per session scan B 8d8feefa3459

Subscribe to this mod's changes

session-orchestrator is a skill published in the GitHub repository rocketmandrey/claude-recall (5 stars, last pushed 2mo ago), licensed MIT. It adds 196 tokens to every session and 3,960 once invoked, about $0.0010 per session on Opus 5. A static security scan graded it B with 1 finding (reads agent configuration directories). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Related

Other skills, from other repositories

project-butler

Project memory workflow for init/upgrade, profile-aware setup, end session, normal/full close, file organization, document archiving, language switching, versioned update logs, rule review, status, wiki sync, and context recovery. Use for /project-butler, setup/初始化, foundation setup, profile setup, end session/收工…

JamesShi96/project-butler · 133 tokens

sync

Fast save-point: sync all session-context files and MEMORY.md with current progress. Zero questions, zero delay. Use when user says /sync, save progress, save state, sync context, or /sync --full for capability inventory.

anombyte93/atlas-session-lifecycle · 49 tokens

memory-quality

Audits and cleans up Claude Code auto-saved memory files. Scores on 4 dimensions, detects conflicts, opens a visual dashboard. Trigger on: memories, memory quality, memory cleanup, memory health, memory dashboard, 检查记忆, 清理记忆, 记忆质量, 记忆看板.

ladyiceberg/memory-quality-mcp · 69 tokens

dataecho-memory

Persistent memory for agents — durable across sessions, machines, sandboxes, and agent platforms, stored in a private DataEcho cloud drive. Use at the start of a session to recall who the user is and what you were working on; whenever the user says "remember this", "don't forget", "like last time", or references past…

mohocp/dataecho · 103 tokens

archive

Persistent knowledge wiki and memory system that compounds across conversations using an Obsidian-compatible vault. Consult at the START of any non-trivial task to recall project context, preferences, and past discoveries. Ingest sources and build wiki pages when accumulating knowledge on a topic. Save at the END when…

jcordon5/archive-skill · 178 tokens

ccthread

Read, search, and summarize Claude Code conversation history. Use when the user asks to find things discussed in past sessions, review old threads, summarize recent work for teammates, extract learnings into a skill or CLAUDE.md, or answer "what was that value/port/command we used?" — ccthread reads /.claude/projects/…

jakemarsh/ccthread · 123 tokens