edheltzel · edheltzel · Jul 22, 2026 · Jul 22, 2026
diff --git a/.agents/plugins/marketplace.json b/.agents/plugins/marketplace.json
@@ -0,0 +1,20 @@
+{
+  "name": "recall-marketplace",
+  "interface": {
+    "displayName": "Recall"
+  },
+  "plugins": [
+    {
+      "name": "recall",
+      "source": {
+        "source": "local",
+        "path": "./plugins/recall"
+      },
+      "policy": {
+        "installation": "AVAILABLE",
+        "authentication": "ON_INSTALL"
+      },
+      "category": "Productivity"
+    }
+  ]
+}
diff --git a/.gitignore b/.gitignore
@@ -18,6 +18,11 @@ dist/
 .claude/worktrees
 .maestro/
 .agents/
+!.agents/
+.agents/*
+!.agents/plugins/
+.agents/plugins/*
+!.agents/plugins/marketplace.json
 .pi-subagents/
 .pi/*
 !.pi/APPEND_SYSTEM.md

diff --git a/AGENTS.md b/AGENTS.md
@@ -26,6 +26,7 @@ Top-level directories, by purpose (one line each — not a file enumeration):
 - `tests/` — `bun:test` suite mirroring source areas, plus install-lifecycle tests
 - `benchmarks/` — wake-up context-efficiency benchmark harness
 - `agent-skills/` — Agent Skills (SKILL.md, one per skill dir) installed to `~/.claude/skills`, `~/.pi/agent/skills`, `~/.omp/agent/skills` — the single `recall-*` command surface (the former `/Recall:*` slash commands, #228)
+- `plugins/` — native host plugin bundles; Codex packages MCP plus generated host-adapted skills in `plugins/recall/`
 - `docs/` — user-facing published docs + ADRs (`docs/adr/`) + agent skill docs (`docs/agents/`)
 - `lib/` — shared bash for the install / update / uninstall lifecycle scripts
 - `opencode/` — OpenCode host integration (plugins / hooks / guide)
@@ -174,5 +175,13 @@ Child AGENTS.md files own domain-specific local rules. Read the applicable one b
 - [`benchmarks/AGENTS.md`](benchmarks/AGENTS.md) — wake-up context-efficiency benchmark harness
 - [`docs/AGENTS.md`](docs/AGENTS.md) — user-facing published docs, ADRs, agent skill docs (never plans/specs)
 - [`agent-skills/AGENTS.md`](agent-skills/AGENTS.md) — `recall-*` Agent Skill definitions
+- [`plugins/AGENTS.md`](plugins/AGENTS.md) — native host plugin manifests, MCP registration, and generated adapters
 
 Owned at root (no child doc): lifecycle scripts (`install.sh`, `update.sh`, `uninstall.sh`) + their shared `lib/install-lib.sh`; platform guides (`FOR_CLAUDE.md`, `FOR_OPENCODE.md`, `FOR_PI.md`); `CONTEXT.md`; and `assets/` (banner + VHS demo tapes/gifs).
+
+## Maintaining this file
+
+Keep this file for knowledge useful to almost every future agent session in this project.
+Do not repeat what the codebase already shows; point to the authoritative file or command instead.
+Prefer rewriting or pruning existing entries over appending new ones.
+When updating this file, preserve this bar for all agents and keep entries concise.
diff --git a/README.md b/README.md
@@ -7,7 +7,7 @@
 Recall is a retrieval-first memory layer: everything lands in one searchable database, the best of it is ranked and injected at session start, and decisions carry confidence, importance, and a lifecycle across any coding agent/harness.
 
 
-> **A SQLite-backed persistent memory layer for coding agents.** Stop-hook extraction captures sessions as you work, MCP tools expose them mid-session, hybrid search (FTS5 + embeddings) retrieves them, and a tiered L0/L1 recall block injects identity + top-ranked records at every session start. Works across Claude Code, OpenCode, and Pi from one local database.
+> **A SQLite-backed persistent memory layer for coding agents.** Stop-hook extraction captures sessions where a host lifecycle adapter exists, MCP tools expose them mid-session, hybrid search (FTS5 + embeddings) retrieves them, and a tiered L0/L1 recall block injects identity + top-ranked records on supported hosts. Works across Claude Code, OpenCode, Pi, and Codex from one local database.
 
 Got questions about the project? I'd suggest using [DeepWiki](https://deepwiki.com/edheltzel/Recall) from Devin/Cognition to ask questions about the project.
 
@@ -18,7 +18,7 @@ All coding agents forget when a session ends. Recall doesn't — it extracts, in
 
 Built on the [Model Context Protocol](https://modelcontextprotocol.io). One SQLite file. No phone-home. No vendor lock-in.
 
-> Stable on [Claude Code](https://claude.com/claude-code). Beta on [Pi](https://pi.dev/) and Alpha for [OpenCode](https://opencode.ai/) (MCP works; lifecycle extensions are early). [Codex CLI](https://github.com/openai/codex) and [Gemini CLI](https://github.com/google-gemini/gemini-cli) on the roadmap. See [Roadmap](#roadmap).
+> Stable on [Claude Code](https://claude.com/claude-code). Beta on [Pi](https://pi.dev/) and Alpha for [OpenCode](https://opencode.ai/) (MCP works; lifecycle extensions are early). [Codex CLI](https://github.com/openai/codex) uses a native plugin for MCP and skills; lifecycle auto-capture is not yet supported. [Gemini CLI](https://github.com/google-gemini/gemini-cli) remains on the roadmap. See [Roadmap](#roadmap).
 
 ---
 
@@ -51,7 +51,7 @@ Install once, then forget about it. Recall runs silently in the background:
 Four things that set Recall apart from cloud-hosted memory layers and from agent-specific scratch files:
 
 - **Local-first, zero infrastructure.** One SQLite file at `~/.agents/Recall/recall.db` (override via `RECALL_DB_PATH`). WAL mode, `0600` perms. No vector database, no graph database, no agent server, no API keys for retrieval. Nothing leaves your machine — no telemetry, no phone-home. Optional Ollama for embeddings (also local).
-- **Multi-agent native.** One memory layer across the agents you actually use. Stable on Claude Code today; Pi and OpenCode connect via MCP; Codex CLI and Gemini CLI on the way. Memories captured by one agent are searchable from any other agent on the same machine.
+- **Multi-agent native.** One memory layer across the agents you actually use. Stable on Claude Code today; Pi, OpenCode, and Codex connect through MCP. Memories captured by one agent are searchable from any other agent on the same machine.
 - **Structured taxonomy, not a flat blob.** Decisions (with supersede/revert lifecycle and confidence scoring), learnings, breadcrumbs, and curated **Library of Alexandria** entries — each has a purpose and a query path. Importance scoring (1–10) surfaces what matters first.
 - **Hybrid search that works offline.** FTS5 keyword search ships with SQLite — no embedding infrastructure required to find anything. Optional Ollama embeddings layer on top for semantic queries. Both are merged via Reciprocal Rank Fusion. Lose Ollama, lose nothing — the keyword path keeps working.
 
@@ -90,6 +90,8 @@ recall doctor       # Health check
 
 Restart your agent (Claude Code, Pi, or OpenCode) to load the MCP server and hooks.
 
+Codex uses its native plugin marketplace instead of the lifecycle installer; see [Codex Integration](docs/CODEX_INTEGRATION.md).
+
 ### First run: set your identity
 
 Recall's tiered RecallStart injects a small identity file at the top
@@ -217,7 +219,7 @@ The source `.excalidraw` file lives at [`assets/how-recall-works.excalidraw`](as
 4. **Extraction pipeline** — The conversation JSONL is filtered, deduplicated, and sent to the `claude` CLI running Haiku (with chunking for large sessions >120K chars). Optional Ollama fallback if the CLI fails. A quality gate rejects low-quality extractions before they're stored.
 5. **PreCompact flush** — When Claude Code is about to compact its context, a `PreCompact` hook (`RecallPreCompact.ts`) flushes the in-flight messages first, so the squashed window is never lost.
 6. **Dual-write storage** — Results are written to SQLite (the only query surface — every CLI/MCP read hits this) and to markdown artifacts (`DISTILLED.md`, `HOT_RECALL.md`, etc., write-only, human-readable).
-7. **Batch catchup (optional)** — A cron job (`RecallBatchExtract.ts`) sweeps any sessions the Stop hook missed during crashes or interruptions, and ingests sessions dropped by the OpenCode plugin and Pi extension into `~/.claude/MEMORY/{opencode,pi}-sessions/`. `install.sh` prints the registration command at the end — opt in by running it once; nothing is auto-scheduled.
+7. **Batch catchup (optional)** — A cron job (`RecallBatchExtract.ts`) sweeps any sessions the Stop hook missed during crashes or interruptions, and ingests sessions dropped by the OpenCode plugin and Pi extension into `~/.agents/Recall/MEMORY/{opencode,pi}-sessions/`. `install.sh` prints the registration command at the end — opt in by running it once; nothing is auto-scheduled.
 8. **TELOS auto-sync (PAI users)** — If you use [Personal AI Infrastructure (PAI)](https://github.com/danielmiessler/Personal_AI_Infrastructure), Recall ships a `RecallTelosSync.ts` SessionStart hook that watches `~/.claude/skills/PAI/USER/TELOS/` for changes and silently runs `recall telos import --update` when any file is newer than the last import. This is **automatic** — no action required once Recall is installed and PAI's TELOS directory exists. You can also import manually at any time with `recall telos import --yes`. If you don't use PAI, the hook checks for the directory, finds nothing, and exits in under 1ms.
 
 ### Search Strategies
@@ -246,7 +248,7 @@ The source `.excalidraw` file lives at [`assets/how-recall-works.excalidraw`](as
 - **Importance scoring (1–10)** — every record carries an importance score that drives what surfaces in L1. Manage with `recall pin` / `recall unpin` / `recall importance backfill`
 - **PreCompact flush** — `RecallPreCompact.ts` writes in-flight messages to SQLite before Claude compacts its context window, so the squashed chunk is never lost
 - **Decision lifecycle** — `recall decision supersede/revert` tracks when a decision was replaced or rolled back; confidence scoring (high/medium/low) on every decision and learning
-- **Cross-host ingestion** — OpenCode plugin and Pi extension drop sessions into `~/.claude/MEMORY/{opencode,pi}-sessions/`; RecallBatchExtract pulls them into the same SQLite DB. One memory layer across agents
+- **Cross-host ingestion** — OpenCode plugin and Pi extension drop sessions into `~/.agents/Recall/MEMORY/{opencode,pi}-sessions/`; RecallBatchExtract pulls them into the same SQLite DB. One memory layer across agents
 - **Library of Alexandria** — curated knowledge entries (session distillations, imported docs, telos goals, quotes) with Fabric `extract_wisdom` analysis. Default importance 8 — these get reserved L1 slots
 - **TELOS integration ([PAI](https://github.com/danielmiessler/Personal_AI_Infrastructure) users)** — `RecallTelosSync.ts` auto-imports your TELOS framework files (goals, mission, projects, strategies) from PAI's `USER/TELOS/` directory on every session start. Changes are detected by mtime; unchanged files are skipped. Manual import: `recall telos import --yes`
 - **Breadcrumbs, decisions, learnings** — three structured record types for non-session memory, addable from CLI (`recall add`), MCP (`memory_add`), or the `recall-add` agent skill
@@ -305,6 +307,7 @@ If you're an AI agent reading this repository:
 | **Using Recall from Claude Code** (MCP tools, CLI, core rules) | [`FOR_CLAUDE.md`](FOR_CLAUDE.md)     |
 | **Using Recall from OpenCode**                                 | [`FOR_OPENCODE.md`](FOR_OPENCODE.md) |
 | **Using Recall from Pi**                                       | [`FOR_PI.md`](FOR_PI.md)             |
+| **Using Recall from Codex**                                    | [`docs/CODEX_INTEGRATION.md`](docs/CODEX_INTEGRATION.md) |
 | **Developing Recall** (build, test, conventions)               | [`CLAUDE.md`](CLAUDE.md)             |
 
 ## Roadmap
@@ -316,7 +319,7 @@ Recall is built around two integration surfaces: **MCP** (memory search and add,
 | [**Claude Code**](https://claude.com/claude-code)             | ✅  |            ✅ Stop · SessionStart · PreCompact             | **Stable** — reference implementation |
 | [**Pi**](https://pi.dev/)                                     | ✅  | ⚠ Beta — `recall-compaction` + `recall-extract` extensions | In progress                           |
 | [**OpenCode**](https://opencode.ai/)                          | ✅  |             ⚠ Alpha — `recall-extract` plugin              | In progress                           |
-| [**Codex CLI**](https://github.com/openai/codex)              |  —  |                             —                              | Coming soon                           |
+| [**Codex CLI**](https://github.com/openai/codex)              | ✅  |               — native plugin, explicit dump only          | MCP + skills available                |
 | [**Gemini CLI**](https://github.com/google-gemini/gemini-cli) |  —  |                             —                              | Coming soon                           |
 
 **Candidate** — [Cursor](https://cursor.com): both `.cursor/hooks.json` and MCP are first-class; the integration model maps cleanly onto Recall's existing hook architecture. Tracked but not started.
@@ -334,6 +337,7 @@ Have an agent you'd like to see supported? [Open an issue](https://github.com/ed
 | [Architecture](docs/architecture.md)       | Database, search, extraction pipeline                                     |
 | Codebase Map (local)                       | Interactive visual map at `.agents/atlas/artifacts/2026-06-10-recall-codebase-map.html` — generated from the codegraph index, not committed (`.agents/` is gitignored) |
 | [Agent Skills](docs/agent-skills.md)       | `recall-*` skills for Claude Code, Pi, and omp                            |
+| [Codex Integration](docs/CODEX_INTEGRATION.md) | Native plugin install, MCP coverage, and lifecycle limits             |
 | [Upgrading](docs/upgrading.md)             | Update, backup, migration system                                          |
 | [Troubleshooting](docs/troubleshooting.md) | Common issues and fixes                                                   |
 | [Changelog](CHANGELOG.md)                  | Release notes and breaking changes                                        |

diff --git a/agent-skills/AGENTS.md b/agent-skills/AGENTS.md
@@ -4,7 +4,7 @@
 
 ## Purpose
 
-Agent Skill definitions (one `SKILL.md` per skill directory, per the [agentskills.io](https://agentskills.io) standard) for the `recall-*` namespace — the single command surface across all skill hosts. The installer symlinks canonicals into `~/.claude/skills/`, `~/.pi/agent/skills/`, and `~/.omp/agent/skills/`.
+Agent Skill definitions (one `SKILL.md` per skill directory, per the [agentskills.io](https://agentskills.io) standard) for the `recall-*` namespace — the canonical command source across skill hosts. The installer symlinks canonicals into `~/.claude/skills/`, `~/.pi/agent/skills/`, and `~/.omp/agent/skills/`; `scripts/build-codex-plugin.ts` generates Codex-native adapters from them.
 
 ## Ownership
 
@@ -14,17 +14,18 @@ Agent Skill definitions (one `SKILL.md` per skill directory, per the [agentskill
 
 - These are Markdown skill specs, not code; each maps to underlying `recall` CLI / MCP behavior — keep its body aligned with that command and with `docs/agent-skills.md`.
 - `recall-dump/SKILL.md` carries `disable-model-invocation: true` — dumping a session is always the user's call; do not remove that gate.
+- Codex does not accept every canonical frontmatter key. Keep host adaptations in `scripts/build-codex-plugin.ts` and verify the generated `plugins/recall/skills/` behavior; never infer parity from matching bytes.
 - `recall-scout/SKILL.md` output is chat-only by default; a persisted scout artifact goes to `.agents/atlas/artifacts/` (see root `AGENTS.md`), never under `agent-skills/` or `docs/`.
 - `recall-scout/SKILL.md` grounds its repo map / key paths / risks in **CodeGraph (primary)** via the external `codegraph` CLI: a `codegraph status --json` capability probe (on an unindexed repo scout **offers** `codegraph init` and runs it only on an explicit user yes — never auto-runs), a cheap orientation bundle (`status --json` + `files --max-depth 2 --no-metadata`), and at most two narrow `explore` calls under the query discipline pinned in the skill, with a grep/tree-walk backstop. CodeGraph is an enhancement, never a hard dependency — scout degrades gracefully when it isn't present.
 - The former `/Recall:*` slash commands (`commands/Recall/`) were retired in favor of these skills (#228); `install.sh`/`update.sh` clean up their symlinks. Do not reintroduce a parallel command surface.
 
 ## Work Guidance
 
-- Add a skill: create `<name>/SKILL.md`, document it in `docs/agent-skills.md`, and add `<name>` to `RECALL_SKILL_NAMES` in `uninstall.sh`.
+- Add a skill: create `<name>/SKILL.md`, document it in `docs/agent-skills.md`, add `<name>` to `RECALL_SKILL_NAMES` in `uninstall.sh`, and regenerate the Codex plugin adapters.
 
 ## Verification
 
-`tests/install/skills.test.ts` (install/uninstall lifecycle), `tests/commands/scout-workflow.test.ts` (scout workflow contracts).
+`tests/install/skills.test.ts` (install/uninstall lifecycle), `tests/commands/scout-workflow.test.ts` (scout workflow contracts), `tests/plugins/codex-plugin.test.ts` (generated Codex adapters).
 
 ## Child DOX Index
 

diff --git a/docs/AGENTS.md b/docs/AGENTS.md
@@ -8,7 +8,7 @@ User-facing published documentation: installation, CLI / MCP / agent-skill refer
 
 ## Ownership
 
-- Reference & guides — `installation.md`, `cli-reference.md`, `mcp-tools.md`, `agent-skills.md`, `architecture.md`, `troubleshooting.md`, `releasing.md`, `upgrading.md`, `OPENCODE_INTEGRATION.md`, `PI_INTEGRATION.md`
+- Reference & guides — `installation.md`, `cli-reference.md`, `mcp-tools.md`, `agent-skills.md`, `architecture.md`, `troubleshooting.md`, `releasing.md`, `upgrading.md`, `OPENCODE_INTEGRATION.md`, `PI_INTEGRATION.md`, `CODEX_INTEGRATION.md`
 - `adr/` — architectural decision records
 - `agents/` — agent skill docs (`issue-tracker.md`, `triage-labels.md`, `board-status.md`, `domain.md`, `worker-flow.md`, `dox-framework.md`)
 
@@ -17,7 +17,7 @@ User-facing published documentation: installation, CLI / MCP / agent-skill refer
 - `docs/` is EXCLUSIVELY user-facing published docs. NEVER store plans, specs, designs, handoffs, or scout artifacts here — those live under `.agents/atlas/` (see root `AGENTS.md`).
 - ADRs are numbered and append-only: change a decision by adding a new `NNNN-*.md`; don't rewrite a past ADR's decision.
 - Keep docs in sync with behavior — a command, MCP, or lifecycle change must update the matching reference (`cli-reference`, `mcp-tools`, `agent-skills`, `installation`, `upgrading`).
-- Host integration guides stay aligned with their root counterparts (`FOR_OPENCODE.md`, `FOR_PI.md`).
+- Host integration guides stay aligned with their root counterparts (`FOR_OPENCODE.md`, `FOR_PI.md`). Codex's canonical guide is `CODEX_INTEGRATION.md` because its native plugin is repository-distributed rather than lifecycle-installed.
 
 ## Work Guidance