tommy0103/obelisk: Every past session, subagent, and workflow -- queryable by your agent, browsable by you

Two sides of the same index

Obelisk has two sides that share one SQLite index:

Agent side — the obelisk CLI owns the local runtime, while a separate agent skill teaches coding agents how to search and query their session history. The agent writes JS queries, runs them locally, and answers in plain language.

App side — an Electron desktop app for humans to browse sessions, manage memories, view usage stats, and see weekly recap cards.

Both read from the same ~/.obelisk/obelisk.sqlite database. The indexer reads Claude Code transcripts from ~/.claude/projects, Codex transcripts from ~/.codex/sessions and ~/.codex/archived_sessions, Kimi Code sessions from ~/.kimi-code/sessions (or $KIMI_CODE_HOME/sessions), Pi sessions from ~/.pi/agent/sessions, and DeepSeek Harness sessions from ~/.dsh/sessions (or $DSH_HOME/sessions).

Multi-provider support

Obelisk indexes every provider into the same SQLite schema instead of keeping separate databases. Rows carry a source value, and non-Claude IDs are provider-prefixed so they cannot collide.

Codex root threads become normal Obelisk sessions. Codex child threads are attached through the same subagents table when parent-thread metadata is available. Codex does not emit Claude-style workflow metadata, so workflow tables may be empty for Codex-only history.

Kimi session directories become one Obelisk session each. Main and child-agent wire.jsonl streams are projected into the same messages, tools, summaries and subagents tables. Undo/clear is handled as a full session replay, so retracted wire records do not remain in the index.

Pi JSONL v1-v3 sessions are projected through the same provider contract. Pi's tree, branch summaries, compactions, durable leaf, retained checkpoint tail, custom messages, bash records, tool calls, token usage, and raw JSONL evidence stay inside the adapter; no Pi-specific database or renderer branch is needed. Active visibility follows Pi's own context rules: a retained tail replaces pre-compaction ancestors even when those physical entries still exist and bounds any later legacy compaction, while a legacy-only chain retains ancestors beginning at firstKeptEntryId. Missing parents form orphan branch roots, matching Pi's recovery behavior. Pi entries that the source explicitly superseded are stored as inactive: the app and normal agent queries omit them, while supported query helpers can include them with includeInactive: true. Display-suppressed or transport-only records remain hidden and are never returned by those helpers.

| Provider | Superseded-history support | | --- | --- | | Pi | Branch, leaf, and compaction state attests inactive history | | Kimi Code | Undo/clear can attest supersession; preservation is a follow-up | | Claude Code | The source does not attest rewind or current-leaf state | | Codex | Sessions have no branching semantics |

Because Pi's explicit --session-id is project-local, Obelisk combines the header ID with a deterministic hash of the normalized header cwd; this keeps the identity stable across file moves and v1-v3 migration while allowing two projects to use the same custom ID. Replacement and deletion replay is provenance-aware, so stale session snapshots are retracted atomically; compaction and branch-summary model usage is included in usage totals.

For live app refresh, Obelisk watches the roots declared by every registered provider, including ~/.claude/projects, ~/.codex/sessions, ~/.codex/archived_sessions, ~/.kimi-code/sessions, and ~/.pi/agent/sessions. Codex's session_index.jsonl is used as lightweight title/update metadata during indexing, not as the message transcript source.

Pi chooses its session directory in this order: --session-dir, PI_CODING_AGENT_SESSION_DIR, sessionDir in settings, then the default under ~/.pi/agent/sessions. Obelisk automatically follows absolute or ~-prefixed environment/global settings and the project setting for Obelisk's launch cwd; a relative project setting is resolved against that cwd. CLI-only roots, relative environment/global settings, and project settings from another launch cwd cannot be inferred safely, so select the resolved directory in Obelisk Settings instead of letting Obelisk guess.

Skill: agent-first retrieval

You can use obelisk like:

/obelisk 上次 auth bug 最后到底改了哪些文件，为什么这么改
/obelisk 这个文件最近在哪些 sessions 里被反复修改
/obelisk 找出最近失败的 tool calls，它们分别发生在哪些任务里
/obelisk 那个 review workflow 的 subagents 各自结论是什么
/obelisk recap this week

Install

Let your agent install it (recommended)

The shortest path is to give the bootstrap guide directly to a coding agent with shell access. Paste this as a prompt into Claude Code, Codex, or another agent — not into your terminal:

Install Obelisk by fetching and following this guide:
curl -fsSL https://raw.githubusercontent.com/tommy0103/obelisk/main/SKILL.md

The agent will ask before changing your machine, install and verify the CLI, then ask whether the formal /obelisk skill should be installed for the current project or globally. The bootstrap guide is only for one-time setup; it is not the query skill itself.

Install manually

Obelisk requires Node.js 22.13 or newer. Install the platform-neutral CLI:

npm install --global @obelisk-apps/cli
obelisk --version

On macOS, Linux, or WSL, the CLI-only installer is equivalent:

curl -fsSL https://raw.githubusercontent.com/tommy0103/obelisk/main/install.sh | sh

Then install the agent skill:

obelisk install

obelisk install delegates to the standard skills installer for tommy0103/obelisk-skill.

Then in any Claude Code session:

/obelisk <your question>

First run builds the index (~5 seconds for 100 sessions). After that it rebuilds incrementally.

How it works

You ask a question
  ↓
Agent writes a JS query against the SQLite index
  ↓
Runs it via obelisk --query <script>
  ↓
Reads the JSON result, answers in natural language

Core API: search(), context(), sql(), plus structured helpers (sessions, memories, summaries, workflows, failures, fileHistory, etc).

Memory layer