### CHANGELOG

# Changelog

All notable changes to this project will be documented in this file.

## [Unreleased]

## [1.6.2] - 2026-07-09

### Changed

- Raised the autoresearch auto-resume ceiling from 20 to 200 turns, while adding a stuck-loop override that stops auto-resume after more than 20 consecutive `discard` or `crash` results in the current segment.

## [1.6.1] - 2026-07-02

### Fixed

- Redirected `workingDir` logs no longer auto-activate autoresearch in unrelated pi sessions.
- `/autoresearch off` now persists across `/tree`, compaction, and reloads: a manual off is recorded as a session activation decision and is no longer overridden just because `log.jsonl` still exists.

## [1.6.0] - 2026-06-08

### Changed

- Autoresearch now stores session files under the `.auto/` subfolder by default, with legacy file fallback for existing sessions.
- The dashboard widget is now always expanded — the full results table renders inline above the editor at all times. Removed the collapsed one-liner mode and the `Ctrl+Shift+T` expand/collapse toggle (and its `shortcuts.toggleDashboard` config key). Fullscreen (`Ctrl+Shift+F`) remains the only dashboard toggle.
- Migrated Pi package imports and dependencies from the `@mariozechner` npm scope to `@earendil-works`.

## [1.5.0] - 2026-06-04

### Changed

- The `init_experiment`, `run_experiment`, and `log_experiment` tools are now revealed to the agent only while autoresearch mode is active, instead of being callable in every session. Outside autoresearch mode the tools are absent from the LLM's schema and system prompt, so the agent can no longer self-start a research loop — entry is via `/autoresearch` or resuming a session with an existing `autoresearch.jsonl`.

## [1.4.0] - 2026-05-06

### Added

- Configurable dashboard keyboard shortcuts. Users can now override or disable the toggle and fullscreen shortcuts with a profile-aware `<agent-dir>/extensions/pi-autoresearch.json` config file, helping autoresearch coexist with other pi extensions that bind the same keys.
- Shortcut resolution tests covering defaults, overrides, disabled shortcuts, partial configs, malformed configs, and extension registration.

### Changed

- Dashboard hints and README documentation now reflect the effective shortcuts from config.

## [1.3.0] - 2026-04-29

### Added

- Deterministic compaction summary. When pi compacts context, autoresearch now bypasses the LLM summarization and injects a lossless markdown summary built from persisted state (experiment rules, ideas backlog, and last 50 runs with ASI fields). This eliminates information loss across compaction boundaries.
- Recent-run deltas in the compaction summary use the full segment baseline, not just the first visible run in the window — percentages stay accurate even for long sessions.
- New test coverage for compaction summary assembly, empty state, re-init segments, 50-run cap, and hidden-baseline delta correctness.

### Fixed

- Post-turn auto-resume no longer tells the agent "don't re-read files" when no compaction happened. Split into two resume messages: a generic one for normal turns and a compaction-specific one that correctly references the summary.

## [1.2.0] - 2026-04-28

### Changed

- Long-running loops now ride pi's auto-compaction instead of stopping. When pi summarizes older messages on context overflow, autoresearch detects the resulting idle and re-prompts the agent to re-read `autoresearch.md`, the tail of `autoresearch.jsonl`, `autoresearch.ideas.md`, and `git log` before continuing.

### Fixed

- Manual `/compact` mid-iteration no longer leaves the loop stuck. `session_compact` now schedules a fresh resume even when no `agent_end` fired for the interrupted turn (so no `pendingResumeMessage` was waiting to be rescheduled). Same fix covers split-turn auto-compactions.
- Compaction during agent setup (before the first `log_experiment`) now resumes. The post-turn gate still requires an experiment this turn to avoid resuming on plain chat replies, but the post-compaction gate is permissive — compaction itself is evidence the loop should continue.
- Rapid back-to-back compactions all resume. Dropped the 5-minute auto-resume cooldown that was sized for a different threat model (chat-only `agent_end` loops); the experiment-this-turn gate plus `MAX_AUTORESUME_TURNS = 20` already cover the looping cases the cooldown was guarding against.

### Removed

- Removed the next-iteration token-cost prediction and its `isContextExhausted` guard — pi's auto-compaction handles overflow, so autoresearch no longer needs to estimate or stop early.
- Removed the `iterationTokens` field from `ExperimentResult` and `autoresearch.jsonl`. Existing log files remain readable; the field is simply ignored. The `token-budget.sh` hook example, which relied on it, has been dropped.
- Removed the never-shipped `autoCompactResume` config option (it was opt-in for an earlier draft of this change).

## [1.1.1] - 2026-04-28

### Added

- Published to the npm registry. Install with `pi install npm:pi-autoresearch`.
- Releases now publish automatically from GitHub Actions via npm trusted publisher (OIDC) with provenance attestation.

## [1.1.0] - 2026-04-24

### Added

- Added optional `autoresearch.hooks/before.sh` and `autoresearch.hooks/after.sh` lifecycle hooks for prospective and retrospective iteration automation.
- Added the `autoresearch-hooks` skill plus example hook scripts for research fetching, learnings capture, notifications, anti-thrash, and idea rotation.

## [1.0.1] - 2026-04-22

### Fixed

- Updated the default dashboard shortcuts to `Ctrl+Shift+T` (toggle) and `Ctrl+Shift+F` (fullscreen).
- Avoided the shortcut conflict with Pi's built-in `Ctrl+X` binding introduced in newer Pi releases.

## [1.0.0] - 2026-04-20

### Added

- Initial stable release of `pi-autoresearch`.

---

### README

<div align="center">
<img  height="120" alt="result" src="https://github.com/user-attachments/assets/c66cbd02-4491-4833-a63a-142cfd7530c1" />

# pi-autoresearch
### Autonomous experiment loop for pi
**[Install](#install)** · **[Usage](#usage)** · **[How it works](#how-it-works)**

</div>

*Try an idea, measure it, keep what works, discard what doesn't, repeat forever.*

An extension for **[pi](https://pi.dev/)** — an AI coding agent that runs in your terminal. pi-autoresearch gives pi the tools and workflow to run autonomous optimization loops: try an idea, benchmark it, keep improvements, revert regressions, repeat.

Inspired by [karpathy/autoresearch](https://github.com/karpathy/autoresearch). Works for any optimization target: test speed, bundle size, LLM training, build times, Lighthouse scores.

---

<img width="1736" height="518" alt="pi-autoresearch" src="https://github.com/user-attachments/assets/5078aa31-3530-406a-85fc-bdeff98679a6" />

---

## Quick start

```bash
pi install npm:pi-autoresearch
```

## What's included

| | |
|---|---|
| **Extension** | Tools + live widget + `/autoresearch` dashboard |
| **Skill** | Gathers what to optimize, writes session files, starts the loop |

### Extension tools

| Tool | Description |
|------|-------------|
| `init_experiment` | One-time session config — name, metric, unit, direction |
| `run_experiment` | Runs any command, times wall-clock duration, captures output |
| `log_experiment` | Records result, auto-commits, updates widget and dashboard |

### `/autoresearch` command

| Subcommand | Description |
|------------|-------------|
| `/autoresearch <text>` | Enter autoresearch mode. If `.auto/prompt.md` exists, resumes the loop with `<text>` as context. Otherwise, sets up a new session. |
| `/autoresearch off` | Leave autoresearch mode. Stops auto-resume and clears runtime state but keeps `.auto/log.jsonl` intact. |
| `/autoresearch clear` | Delete `.auto/log.jsonl`, reset all state, and turn autoresearch mode off. Use this for a clean start. |
| `/autoresearch export` | Open a live dashboard in your browser. Auto-updates as experiments run. |

**Examples:**

```
/autoresearch optimize unit test runtime, monitor correctness
/autoresearch model training, run 5 minutes of train.py and note the loss ratio as optimization target
/autoresearch export
/autoresearch off
/autoresearch clear
```

### Keyboard shortcuts

| Shortcut     | Description |
|--------------|-------------|
| `Ctrl+Shift+F` | Open fullscreen scrollable dashboard overlay. Navigate with `↑`/`↓`/`j`/`k`, `PageUp`/`PageDown`/`u`/`d`, `g`/`G` for top/bottom, `Escape` or `q` to close. |

To avoid conflicts with other pi extensions, override or disable these shortcuts in
`<agent-dir>/extensions/pi-autoresearch.json`. `<agent-dir>` is the active pi profile
config directory (usually `~/.pi/agent`, or `PI_CODING_AGENT_DIR` when set):

```json
{
  "shortcuts": {
    "fullscreenDashboard": "ctrl+shift+y"
  }
}
```

Use `null` to skip registering a shortcut. Omitted shortcuts keep their defaults.

### UI

- **Dashboard widget** — always visible above the editor: a full results table with columns for commit, metric, status, and description.
- **Confidence score** — after 3+ runs, shows how the best improvement compares to the session noise floor. ≥2.0× (green) = likely real, 1.0–2.0× (yellow) = above noise but marginal, <1.0× (red) = within noise.
- **Fullscreen overlay** — `Ctrl+Shift+F` opens a scrollable full-terminal dashboard. Shows a live spinner with elapsed time for running experiments.

### Skills

**`autoresearch-create`** asks a few questions (or infers from context) about your goal, command, metric, and files in scope — then writes two files and starts the loop immediately:

**`autoresearch-finalize`** turns a noisy autoresearch branch into clean, independent branches — one per logical change, each starting from the merge-base. Groups must not share files, so each branch can be reviewed and merged independently.

**`autoresearch-hooks`** *(optional)* helps author `.auto/hooks/before.sh` and `.auto/hooks/after.sh` for a session. It ships with ten reference scripts in [`skills/autoresearch-hooks/examples/`](skills/autoresearch-hooks/examples/) (external search, learnings journal, native notifications, anti-thrash, idea rotation, and more) — the skill handles the contract, you pick the inspiration. The core autoresearch loop has no hook awareness.

All session files live in a single `.auto/` subfolder at the working-directory root — one folder to preserve across reverts, gitignore, and clean up. (Legacy flat `autoresearch.*` files are still read for in-flight sessions.)

| File | Purpose |
|------|---------|
| `.auto/prompt.md` | Session document — objective, metrics, files in scope, what's been tried. A fresh agent can resume from this alone. |
| `.auto/measure.sh` | Benchmark script — pre-checks, runs the workload, outputs `METRIC name=number` lines. |
| `.auto/log.jsonl` | Append-only log of every run (written by the tools). |
| `.auto/checks.sh` | *(optional)* Backpressure checks — tests, types, lint. Runs after each passing benchmark. Failures block `keep`. |
| `.auto/hooks/` | *(optional)* Executable scripts (`before.sh`, `after.sh`) that fire around iterations. Stdout is delivered to the agent as a steer message. |

---

## Install

```bash
pi install npm:pi-autoresearch
```

<details>
<summary>Manual install</summary>

```bash
cp -r extensions/pi-autoresearch ~/.pi/agent/extensions/
cp -r skills/autoresearch-create ~/.pi/agent/skills/
```

Then `/reload` in pi.

</details>

---

## Usage

### 1. Start autoresearch

```
/skill:autoresearch-create
```

The agent asks about your goal, command, metric, and files in scope — or infers them from context. It then creates a branch, writes `.auto/prompt.md` and `.auto/measure.sh`, runs the baseline, and starts looping immediately.

### 2. The loop

The agent runs autonomously: edit → commit → `run_experiment` → `log_experiment` → keep or revert → repeat. It never stops unless interrupted.

Every result is appended to `.auto/log.jsonl` in your project — one line per run. This means:

- **Survives restarts** — the agent can resume a session by reading the file
- **Survives context resets** — `.auto/prompt.md` captures what's been tried so a fresh agent has full context
- **Human readable** — open it anytime to see the full history
- **Branch-aware** — each branch has its own session

### 3. Finalize into reviewable branches

```
/skill:autoresearch-finalize
```

The agent reads `.auto/log.jsonl`, groups kept experiments into logical changesets, proposes the grouping for your approval, then creates independent branches from the merge-base. Each commit includes metric improvements in the message. Groups must not share files, so branches can be reviewed and merged independently.

### 4. Monitor progress

- **Widget** — full results table, always visible above the editor
- **`Ctrl+Shift+F`** — fullscreen scrollable dashboard overlay (config key: `shortcuts.fullscreenDashboard`)
- **`/autoresearch export`** — open a live browser dashboard with chart and share card
- **`Escape`** — interrupt anytime and ask for a summary

---

## Example domains

| Domain | Metric | Command |
|--------|--------|---------|
| Test speed | seconds ↓ | `pnpm test` |
| Bundle size | KB ↓ | `pnpm build && du -sb dist` |
| LLM training | val_bpb ↓ | `uv run train.py` |
| Build speed | seconds ↓ | `pnpm build` |
| Lighthouse | perf score ↑ | `lighthouse http://localhost:3000 --output=json` |

---

## How it works

The **extension** is domain-agnostic infrastructure. The **skill** encodes domain knowledge. This separation means one extension serves unlimited domains.

```
┌──────────────────────┐     ┌──────────────────────────┐
│  Extension (global)  │     │  Skill (per-domain)       │
│                      │     │                           │
│  run_experiment      │◄────│  command: pnpm test       │
│  log_experiment      │     │  metric: seconds (lower)  │
│  widget + dashboard  │     │  scope: vitest configs    │
│                      │     │  ideas: pool, parallel…   │
└──────────────────────┘     └──────────────────────────┘
```

Two files keep the session alive across restarts and context resets:

```
.auto/log.jsonl   — append-only log of every run (metric, status, commit, description)
.auto/prompt.md   — living document: objective, what's been tried, dead ends, key wins
```

A fresh agent with no memory can read these two files and continue exactly where the previous session left off.

---

## Configuration (optional)

Create `.auto/config.json` in your pi session directory to customize behavior:

```json
{
  "workingDir": "/path/to/project",
  "maxIterations": 50
}
```

| Field | Type | Description |
|-------|------|-------------|
| `workingDir` | string | Override the directory for all autoresearch operations — file I/O, command execution, and git. Supports absolute or relative paths (resolved against the pi session cwd). The config file itself always stays under the session cwd. Fails if the directory doesn't exist. |
| `maxIterations` | number | Maximum experiments before auto-stopping. The agent is told to stop and won't run more experiments until a new segment is initialized. |

### Long-running loops and context

The loop is designed to run unattended across context limits. When pi's [auto-compaction](https://github.com/badlogic/pi-mono/blob/main/packages/coding-agent/docs/compaction.md) summarizes the older portion of the conversation, autoresearch detects the resulting idle and re-prompts the agent to re-read `.auto/prompt.md`, the tail of `.auto/log.jsonl`, `.auto/ideas.md`, and `git log` before continuing. All progress is persisted in those files, so the post-summary turn rehydrates from the source of truth instead of relying on whatever survived compaction. No tuning required — if pi's auto-compaction is enabled (the default), this just works.

---

## Confidence scoring

After 3+ experiments in a session, pi-autoresearch computes a **confidence score** — how the best improvement compares to the session's noise floor. This helps distinguish real gains from benchmark jitter, especially on noisy signals like ML training, Lighthouse scores, or flaky benchmarks.

**How it works:**

- Uses [Median Absolute Deviation (MAD)](https://en.wikipedia.org/wiki/Median_absolute_deviation) of all metric values in the current segment as a robust noise estimator.
- Confidence = `|best_improvement| / MAD`. A score of 2.0× means the best improvement is twice the noise floor.
- Shown in the widget, fullscreen dashboard, and `log_experiment` output.
- Persisted to `.auto/log.jsonl` on each result for post-hoc analysis.
- **Advisory only** — never auto-discards. The agent is guided to re-run experiments when confidence is low, but the final keep/discard decision stays with the agent.

| Confidence | Color | Meaning |
|-----------|-------|---------|
| ≥ 2.0× | 🟢 green | Improvement is likely real |
| 1.0–2.0× | 🟡 yellow | Above noise but marginal |
| < 1.0× | 🔴 red | Within noise — consider re-running to confirm |

---

## Backpressure checks (optional)

Create `.auto/checks.sh` to run correctness checks (tests, types, lint) after every passing benchmark. This ensures optimizations don't break things.

```bash
#!/bin/bash
set -euo pipefail
pnpm test --run
pnpm typecheck
```

**How it works:**

- If the file doesn't exist, everything behaves exactly as before — no changes to the loop.
- If it exists, it runs automatically after every benchmark that exits 0.
- Checks execution time does **not** affect the primary metric.
- If checks fail, the experiment is logged as `checks_failed` (same behavior as a crash — no commit, revert changes).
- The `checks_failed` status is shown separately in the dashboard so you can distinguish correctness failures from benchmark crashes.
- Checks have a separate timeout (default 300s, configurable via `checks_timeout_seconds` in `run_experiment`).

---

## Hooks (optional)

Drop executable scripts in `.auto/hooks/` to run code at iteration boundaries. Hooks are **transparent to the agent** — the agent calls tools and sees results; hooks run alongside without any agent-facing surface.

- `.auto/hooks/before.sh` — fires before every iteration (at `/autoresearch` activation and at the end of every `log_experiment`, after `after.sh`). Use for prospective work: fetch research, prime context for the next attempt.
- `.auto/hooks/after.sh` — fires at the end of every `log_experiment`. Use for retrospective work: annotate learnings, send notifications.

**Contract:**

- Must be executable (`chmod +x`). Preserved on revert — the entire `.auto/` folder survives (as do legacy `autoresearch.*` artefacts).
- **Stdin** — a JSON object on a single line. Shape depends on the stage (see below). Extract fields with `jq`.
- **Stdout** is delivered to the agent as a steer message (capped at 8 KB). Empty stdout = silent.
- Non-zero exit or >30s timeout surfaces an error steer to the agent.
- Each fire appends a `{"type":"hook",…}` entry to `.auto/log.jsonl` for observability.

**`before.sh` stdin** (on fresh activation `last_run` is `null`):

```json
{
  "event": "before",
  "cwd": "/path/to/workdir",
  "next_run": 6,
  "last_run": {
    "run": 5, "status": "discard", "metric": 42.1,
    "description": "…",
    "asi": { "hypothesis": "…", "next_focus": "…" }
  },
  "session": {
    "metric_name": "total_ms", "metric_unit": "ms", "direction": "lower",
    "baseline_metric": 40.7, "best_metric": 33.5,
    "run_count": 5, "goal": "optimize sort speed"
  }
}
```

**`after.sh` stdin:**

```json
{
  "event": "after",
  "cwd": "/path/to/workdir",
  "run_entry": {
    "run": 6, "status": "discard", "metric": 38.9,
    "description": "…",
    "asi": { "hypothesis": "…", "learned": "…" }
  },
  "session": { "metric_name": "total_ms", "direction": "lower", "baseline_metric": 40.7, "best_metric": 33.5, "run_count": 6, "goal": "…" }
}
```

**Agent signal.** The agent writes `description` and `asi.*` fields in its `log_experiment` calls for its own future-self reasoning. The hook opportunistically mines whichever fields the agent naturally uses — `asi.hypothesis`, `asi.next_focus`, `description`, etc. There is no dedicated "hook input" field; the agent is unaware the hook exists.

**Examples.** Reference scripts for both stages live at [`skills/autoresearch-hooks/examples/`](skills/autoresearch-hooks/examples/) — external search, qmd document search, persistent learnings, native notifications, git tagging, anti-thrash, idea rotator, hypothesis reflection, context rotation. Copy one to your session's `.auto/hooks/` directory, adapt, `chmod +x`.

---

## Prerequisites

1. **Install pi** — follow the instructions at [pi.dev](https://pi.dev/)
2. **An API key** for your preferred LLM provider (configured in pi)

## Controlling costs

Autoresearch loops run autonomously and can burn through tokens. Two ways to cap spend:

- **API key limits** — most providers let you set per-key or monthly budgets. Check your provider's dashboard.
- **`maxIterations`** — cap experiments per session in `.auto/config.json`:
   ```json
   {
     "maxIterations": 30
   }
   ```

## License

MIT

---