---
title: "Tools"
description: "Software Aayush Manchanda installed, ran, and formed an opinion about, with a dated verdict on each one."
canonical: "https://aayushmanchanda.com/tools"
source: "https://aayushmanchanda.com/tools.md"
last-updated: 2026-08-26
---

# Tools

Software Aayush Manchanda installed, ran, and formed an opinion about, with a dated verdict on each one.

Source: https://aayushmanchanda.com/tools

| Tool | Verdict | Category | Updated | Note |
| --- | --- | --- | --- | --- |
| [Paperclip](https://github.com/paperclipai/paperclip) | using | agent infra | 2026-08-17 | Agent org control plane. Verified working on the overnight bench — boots clean, embedded Postgres, telemetry off. |
| [Hermes](https://github.com/NousResearch/hermes-agent) | using | agent infra | 2026-08-26 | Telegram-gateway agent runtime. My daily driver. |
| [Headlong](https://github.com/laude-institute/headlong) | watching | agent infra | 2026-08-26 | Persistent agent harness. Installed on a Factory droid backend; first real run still pending. |
| [Cabinet](https://github.com/cabinetai/cabinet) | on-hold | agent infra | 2026-08-17 | Won my head-to-head demo comparison against two alternatives; full re-scaffold pending. Promising. |
| [cloudflare-os](https://github.com/cloudflare/cloudflare-os) | skipped | agent infra | 2026-08-17 | Lost the same head-to-head to Cabinet on my use case. |
| [Synara](https://github.com/Emanuele-web04/synara) | watching | agent workspaces | 2026-08-19 | Desktop app builds and boots clean end-to-end (Bun + Electron). Works; not yet part of the daily stack. |
| [Buzz](https://github.com/block/buzz) | on-hold | agent workspaces | 2026-08-17 | Block's agent workspace. Cloned, never judged — the source build wants multi-GB toolchain and disk ran out mid-bench. Prebuilt release is the easy path. |
| [WUPHF](https://github.com/nex-crm/wuphf) | skipped | agent workspaces | 2026-08-17 | Weaker take on what Buzz does, with a restrictive license on top. |
| [improve (shadcn skill)](https://github.com/shadcn/improve) | using | claude skills | 2026-08-17 | Plan-cheap-execute optimization loop as a Claude skill. Cherry-picked from the bench. |
| [davidondrej skill pack](https://github.com/davidondrej/skills) | using | claude skills | 2026-08-17 | Three skills kept after a security read: risky-changes (adapted), folder-specific-claude-md, effective-agent-skills. |
| [lazyweb skill](https://lazyweb.com) | skipped | claude skills | 2026-08-17 | Design-research skill pack with a curl\|bash installer; paid tier can decrypt browser cookies. Only ever deliberately, with the cookie-decryption caveat in mind. |
| [agent-browser](https://agent-browser.dev) | using | web utility | 2026-08-26 | Browser automation CLI for agents — accessibility-tree snapshots at a fraction of screenshot token cost. Daily driver. |
| [Firecrawl](https://firecrawl.dev) | using | web utility | 2026-08-26 | Scrape/crawl/extract API for pages that fight plain fetch. Its X postprocessor is the only reliable way I read tweets programmatically. |
| [Is Agentic](https://is-agentic.com/) | watching | web utility | 2026-08-26 | Score how ready a website is for AI agents, then get evidence and recommendations to improve it. |
| [pxpipe](https://github.com/teamchong/pxpipe) | skipped | web utility | 2026-08-17 | Token-saving MITM proxy with documented silent string corruption. Unusable anywhere fidelity matters. |
| [Dinky](https://github.com/heyderekj/dinky) | skipped | desktop utility | 2026-08-17 | Unsigned app from a personal tap — not worth the trust exception. |
| [hubble.md](https://hubble.md) | skipped | desktop utility | 2026-08-17 | Duplicates my existing markdown vault workflow. |
| [Macro](https://macro.com) | skipped | desktop utility | 2026-08-17 | The workspace product turned out to be a Notion replacement, not the PDF tool I wanted. |
| [Papercuts](https://github.com/treygoff24/papercuts) | skipped | desktop utility | 2026-08-17 | Fine tool, but requires a Rust toolchain I don't want to carry right now. |
| [Eve](https://github.com/vercel/eve) | watching | sandboxes | 2026-08-17 | Vercel's agent sandbox. Scaffolded and ready to test — needs Node 24. |

## In my words

Some entries carry more than the one-line note. Most do not, and an entry missing from this list simply means I have not written it up, not that I had nothing good to say.

### agent-browser

- What I like: The accessibility-tree snapshots cost a fraction of what screenshots do. The project claims about 82 percent less; feels right in my use. The tree comes with element refs, so an agent clicks a named thing instead of guessing at pixel coordinates.

### Firecrawl

- Why: When the job is to get the content, not to click through a UI. A page that blocks a plain fetch, a whole site I want as markdown, a structured field I want pulled out. If the job needs a login or real clicking, I reach for agent-browser instead.
- What I like: Its X postprocessor is the only reliable way I read tweets programmatically.
- Try: `npx -y firecrawl-cli@latest`

### pxpipe

- What I don't: It corrupts strings on the way through and says nothing about it. That is not my finding, it is in the project's own documentation. A proxy that quietly changes the bytes is worse than no proxy at all, because you do not find out where it broke, you find out somewhere downstream when a thing that should have matched does not. No token saving buys that back.

## Verdicts

- `using`: In the daily stack right now.
- `watching`: Runs fine and has not earned a place in the daily stack.
- `on-hold`: Testing started, then stopped before reaching a verdict.
- `skipped`: Looked at and set aside. The note says why.

Every verdict carries the date it was last true. How stale that makes it is the reader's call to make, not a thing the page decides.

## Filtered views

By category:

- agent infra (5): https://aayushmanchanda.com/tools/category/agent-infra
- agent workspaces (3): https://aayushmanchanda.com/tools/category/agent-workspaces
- claude skills (3): https://aayushmanchanda.com/tools/category/claude-skills
- web utility (4): https://aayushmanchanda.com/tools/category/web-utility
- desktop utility (4): https://aayushmanchanda.com/tools/category/desktop-utility
- sandboxes (1): https://aayushmanchanda.com/tools/category/sandboxes

By verdict:

- using (6): https://aayushmanchanda.com/tools/verdict/using
- watching (4): https://aayushmanchanda.com/tools/verdict/watching
- on-hold (2): https://aayushmanchanda.com/tools/verdict/on-hold
- skipped (8): https://aayushmanchanda.com/tools/verdict/skipped

## One page per tool

- Paperclip: https://aayushmanchanda.com/tools/paperclip
- Hermes: https://aayushmanchanda.com/tools/hermes
- Headlong: https://aayushmanchanda.com/tools/headlong
- Cabinet: https://aayushmanchanda.com/tools/cabinet
- cloudflare-os: https://aayushmanchanda.com/tools/cloudflare-os
- Synara: https://aayushmanchanda.com/tools/synara
- Buzz: https://aayushmanchanda.com/tools/buzz
- WUPHF: https://aayushmanchanda.com/tools/wuphf
- improve (shadcn skill): https://aayushmanchanda.com/tools/improve-skill
- davidondrej skill pack: https://aayushmanchanda.com/tools/davidondrej-skill-pack
- lazyweb skill: https://aayushmanchanda.com/tools/lazyweb-skill
- agent-browser: https://aayushmanchanda.com/tools/agent-browser
- Firecrawl: https://aayushmanchanda.com/tools/firecrawl
- Is Agentic: https://aayushmanchanda.com/tools/is-agentic
- pxpipe: https://aayushmanchanda.com/tools/pxpipe
- Dinky: https://aayushmanchanda.com/tools/dinky
- hubble.md: https://aayushmanchanda.com/tools/hubble-md
- Macro: https://aayushmanchanda.com/tools/macro
- Papercuts: https://aayushmanchanda.com/tools/papercuts
- Eve: https://aayushmanchanda.com/tools/eve

## Other pages

- Home: https://aayushmanchanda.com/index.md
- Sites: https://aayushmanchanda.com/sites.md
- Reading: https://aayushmanchanda.com/reading.md
- Experiments: https://aayushmanchanda.com/experiments.md
- Notes: https://aayushmanchanda.com/notes.md
- Site summary for agents: https://aayushmanchanda.com/llms.txt
- Every URL the site publishes: https://aayushmanchanda.com/sitemap-index.xml