Add: SLICC

[trieloff]

Summary

Adds SLICC, a browser-based agent runtime, to the bottom of the Autonomous Web Agents section.

Item

• Name: SLICC
• Link: https://www.sliccy.com

Section

Autonomous Web Agents

Why this belongs

SLICC is itself a web agent rather than a tool for building one. It runs inside Chrome and operates the web by driving the user's own tabs: it opens pages, fills and submits forms, extracts data, and carries out multi-step tasks against live web applications from a natural-language instruction.

It fits this section rather than the neighbouring ones because it is a user-facing agent with its own interface, not a library or framework (which would suit AI Web Automation Tools or Dev Tools), and because it acts on web pages rather than on the host operating system's GUI (which would suit Computer-use Agents).

Relative to the existing browser-extension agents already listed here, such as rtrvr.ai, Nanobrowser, Harpa.ai and Dassi, SLICC differs in what the agent can do between web interactions: it carries a POSIX shell with roughly 259 commands, a virtual filesystem, git, and an orchestration model that runs sandboxed sub-agents in parallel with per-agent path and command allow-lists. So a task can interleave browsing with local text processing, version control, and scheduled follow-up in one session.

Primary documented web-agent use case

Web interaction is the primary documented use case, not an incidental one:

• https://www.sliccy.com — the product's own description is automating web apps through the browser's existing sessions.
• https://www.sliccy.com/automate-website-migration — an end-to-end web task: reading a live site and migrating its content.
• https://www.sliccy.com/use-cases/sandbox and https://www.sliccy.com/use-cases/explore — agent-driven browsing and page interaction.
• https://github.com/ai-ecoverse/slicc — the public runtime repository.
• https://www.sliccy.com/man/playwright-cli — reference page for the browser-driving command the agent uses to navigate and act on pages.

Public reference

https://github.com/ai-ecoverse/slicc (public repository, 30 stars) and https://www.sliccy.com (documentation, including a per-command reference under /man/).

Affiliation disclosure

Yes, I am affiliated. I maintain the SLICC website and contribute to its skills repository, so treat this as a first-party submission and weigh the description accordingly.

Two notes in that spirit, since the policy asks for factual and neutral wording:

• SLICC requires the user to supply their own model provider tokens. It removes the need for API keys for the web apps being automated, not for the model.
• The Chrome extension path needs no local install; there is also an npm path that runs a Node process which launches Chrome. The entry says "browser-based" rather than "no install" for that reason.

Checklist

• This PR adds exactly one item.
• I added the item to the bottom of the best-fit section.
• The description is factual, neutral, and ends with a period.
• This is directly relevant to web agents, not just AI tooling in general.
• I checked for duplicates before submitting.

Unrelated observation

While checking for duplicates I noticed two entries whose products have been discontinued, which may be candidates for ARCHIVE.md in a separate PR:

• OpenAI Operator was folded into ChatGPT agent and then into ChatGPT Atlas. OpenAI's own page states Atlas "is scheduled to stop working on August 9, 2026" (https://help.openai.com/en/articles/20001371), a date now past.
• MultiOn no longer resolves to the product; multion.ai serves a different company.

Happy to open a separate Archive: PR for those if useful, or to leave them alone.