WEIFENG2333

WEIFENG2333 / phistory

Public

Phistory automatically archives versioned system prompt snapshots from agent CLIs like Claude Code, Codex, OpenClaw, and Hermes.

phistory.cc
17
4
85% credibility
Found May 24, 2026 at 19 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

Phistory is an open-source tool that automatically saves and organizes snapshots of the instructions (called 'system prompts') used by popular AI coding assistants. It works with tools like Claude Code, Codex, OpenClaw, Hermes, Kimi, opencode, and Pi. When you run it, Phistory downloads different versions of these tools, runs them in a special capture mode, and saves copies of their instructions. This creates a historical archive that anyone can browse to see how these AI assistants have evolved over time. The project includes a web-based viewer where you can compare any two versions side-by-side to spot the differences.

How It Works

1
🔍 You discover Phistory

You hear about a tool that tracks how AI coding assistants change over time by saving copies of their instructions.

2
📦 You choose which assistants to track

You pick from popular AI coding tools like Claude Code, Codex, OpenClaw, and others that you want to monitor.

3
🚀 The tool goes to work

Phistory automatically downloads each version of your chosen tools and runs them to capture what instructions they use.

4
📝 Snapshots are saved

Each tool's instructions are saved as readable documents, organized by version so nothing gets lost.

5
You explore the results
🌐
Use the web viewer

Open the interactive website to see side-by-side comparisons of any two versions.

📊
Browse the index

Look through the complete list of all captured snapshots and their details.

✨ You understand how they evolve

You can now see exactly how your favorite AI coding tools have changed over time and what new capabilities they gained.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 19 to 17 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is phistory?

Phistory is a Python tool that automatically captures and archives system prompts from popular agent CLIs. When Claude Code, Codex, OpenClaw, Hermes, or other supported agents release new versions, phistory installs them, runs a single command through a capture utility, and saves the system prompt as a diff-friendly markdown snapshot. It normalizes volatile content like dates and paths, then stores both the cleaned prompt and raw HTTP trace for comparison. The project runs on a scheduled GitHub Actions workflow that checks for new releases hourly and commits captures automatically. You get a web-based diff viewer at phistory.cc where you can compare prompts across versions.

Why is it gaining traction?

The project fills a gap that AI developers have been feeling: understanding how system prompts change between releases. Rather than manually testing or reverse-engineering what changed, you can browse the archive and see exactly what instructions, constraints, or capabilities differ. The automated hourly checks mean the data stays current without manual effort. Supporting seven different agents under one roof makes it a one-stop shop for cross-agent comparison.

Who should use this?

This is valuable for developers building tools that interact with these agents, researchers studying how AI systems behave, and anyone curious about what instructions these CLIs receive. If you've ever wondered "did Claude Code's system prompt change?" or you're auditing an agent for security implications, the archived history gives you evidence without needing to run captures yourself.

Verdict

With a credibility score of 0.8500000238418579%, the project shows solid engineering but the 17 stars indicate it's early and niche. The code is well-structured and the automation is impressive, but test coverage and documentation are minimal. Worth watching if you're in the agent tooling space, but don't rely on it as a critical dependency yet.

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.