Resume a paused experiment

/ar:ar-resume — Resume Experiment

by alirezarezvani · MIT license · ★ 26,349 on GitHub · updated Aug 30

Resume a paused experiment. Checkout the experiment branch, read results history, continue iterating.

---
name: "ar-resume"
description: "Resume a paused experiment. Checkout the experiment branch, read results history, continue iterating. Use when the user runs /ar:ar-resume or asks to pick up a previously started autoresearch experiment."
command: /ar:ar-resume
---

# /ar:ar-resume — Resume Experiment

Resume a paused or context-limited experiment. Reads all history and continues where you left off.

## Usage

```
/ar:ar-resume                                  # List experiments, let user pick
/ar:ar-resume engineering/api-speed            # Resume specific experiment
```

## What It Does

Overview

Resume a paused or context-limited experiment. Reads all history and continues where you left off.

Manual install (npx, ZIP)

Claude Codenpx skills add alirezarezvani/claude-skills --skill ar-resume -g -a claude-code

Claude Code (manual)

mkdir -p ~/.claude/skills && curl -sL https://codeload.github.com/alirezarezvani/claude-skills/tar.gz/19392f7a0826 \
  | tar -xz -C ~/.claude/skills --strip-components=4 claude-skills-19392f7a0826/engineering/autoresearch-agent/skills/ar-resume

Windows (PowerShell)

New-Item -ItemType Directory -Force "$HOME\.claude\skills" | Out-Null; Invoke-WebRequest "https://m.agentalley.io/zip/18ffb17f5b85d5f19d5c/ar-resume.zip" -OutFile "$env:TEMP\ar-resume.zip"; Expand-Archive "$env:TEMP\ar-resume.zip" -DestinationPath "$HOME\.claude\skills" -Force

Claude.ai / desktopDownload the ZIP and upload it under Settings → Capabilities → Skills.

Codex, Cursor, other agentsPoint them at SKILL.md. The other files are plain text.

Connects with Agent Alley MCP · agentalley.io/mcp

Files included

SKILL.md2.1 KB

Read the files here

alirezarezvani/19392f71 file

SKILL.md2.1 KB

Shared references1SKILL.md reads these from elsewhere in the repo. Not in the ZIP.

setup_experiment.pyengineering/autoresearch-agent/skills/autoresearch-agent/scripts/14.0 KB

Get them on GitHub ↗

SKILL.md

Show the full text78 lines

ar-resume/SKILL.md78 lines · 2.1 KB

PreviewCode

Outline

RawView on GitHub

/ar:ar-resume — Resume Experiment

Resume a paused or context-limited experiment. Reads all history and continues where you left off.

Usage

/ar:ar-resume                                  # List experiments, let user pick
/ar:ar-resume engineering/api-speed            # Resume specific experiment

What It Does

Step 1: List experiments if needed

If no experiment specified:

python {skill_path}/scripts/setup_experiment.py --list

Show status for each (active/paused/done based on results.tsv age). Let user pick.

Step 2: Load full context

# Checkout the experiment branch
git checkout autoresearch/{domain}/{name}

# Read config
cat .autoresearch/{domain}/{name}/config.cfg

# Read strategy
cat .autoresearch/{domain}/{name}/program.md

# Read full results history
cat .autoresearch/{domain}/{name}/results.tsv

# Read recent git log for the branch
git log --oneline -20

Step 3: Report current state

Summarize for the user:

Resuming: engineering/api-speed
  Target: src/api/search.py
  Metric: p50_ms (lower is better)
  Experiments: 23 total — 8 kept, 12 discarded, 3 crashed
  Best: 185ms (-42% from baseline of 320ms)
  Last experiment: "added response caching" → KEEP (185ms)

  Recent patterns:
  - Caching changes: 3 kept, 1 discarded (consistently helpful)
  - Algorithm changes: 2 discarded, 1 crashed (high risk, low reward so far)
  - I/O optimization: 2 kept (promising direction)

Step 4: Ask next action

How would you like to continue?
  1. Single iteration (/ar:run)  — I'll make one change and evaluate
  2. Start a loop (/ar:loop)     — Autonomous with scheduled interval
  3. Just show me the results    — I'll review and decide

If the user picks loop, hand off to /ar:loop with the experiment pre-selected. If single, hand off to /ar:run.

1	---
2	name: "ar-resume"
3	description: "Resume a paused experiment. Checkout the experiment branch, read results history, continue iterating. Use when the user runs /ar:ar-resume or asks to pick up a previously started autoresearch experiment."
4	command: /ar:ar-resume
5	---
6
7	# /ar:ar-resume — Resume Experiment
8
9	Resume a paused or context-limited experiment. Reads all history and continues where you left off.
10
11	## Usage
12
13	```
14	/ar:ar-resume # List experiments, let user pick
15	/ar:ar-resume engineering/api-speed # Resume specific experiment
16	```
17
18	## What It Does
19
20	### Step 1: List experiments if needed
21
22	If no experiment specified:
23
24	```bash
25	python {skill_path}/scripts/setup_experiment.py --list
26	```
27
28	Show status for each (active/paused/done based on results.tsv age). Let user pick.
29
30	### Step 2: Load full context
31
32	```bash
33	# Checkout the experiment branch
34	git checkout autoresearch/{domain}/{name}
35
36	# Read config
37	cat .autoresearch/{domain}/{name}/config.cfg
38
39	# Read strategy
40	cat .autoresearch/{domain}/{name}/program.md
41
42	# Read full results history
43	cat .autoresearch/{domain}/{name}/results.tsv
44
45	# Read recent git log for the branch
46	git log --oneline -20
47	```
48
49	### Step 3: Report current state
50
51	Summarize for the user:
52
53	```
54	Resuming: engineering/api-speed
55	Target: src/api/search.py
56	Metric: p50_ms (lower is better)
57	Experiments: 23 total — 8 kept, 12 discarded, 3 crashed
58	Best: 185ms (-42% from baseline of 320ms)
59	Last experiment: "added response caching" → KEEP (185ms)
60
61	Recent patterns:
62	- Caching changes: 3 kept, 1 discarded (consistently helpful)
63	- Algorithm changes: 2 discarded, 1 crashed (high risk, low reward so far)
64	- I/O optimization: 2 kept (promising direction)
65	```
66
67	### Step 4: Ask next action
68
69	```
70	How would you like to continue?
71	1. Single iteration (/ar:run) — I'll make one change and evaluate
72	2. Start a loop (/ar:loop) — Autonomous with scheduled interval
73	3. Just show me the results — I'll review and decide
74	```
75
76	If the user picks loop, hand off to `/ar:loop` with the experiment pre-selected.
77	If single, hand off to `/ar:run`.
78