Hacker ApprovedDepartments

smol

Shares rank 6 of 11 coding agent harnesses by recommenders

A minimal open source coding agent that is small, light on CPU and RAM, and easy to read and adapt. Critics note it lacks subagents and a system prompt, and its size hurts caching.

A minimal open source coding agent that is small, light on CPU and RAM, and easy to read and adapt. Critics note it lacks subagents and a system prompt, and its size hurts caching.

10 people gave an opinion

recommend it
9
criticize it
1
approval share
90%
Each figure is one person. RecommendsCriticizes

Recommended in 5 of 13 months

  1. November 2025: recommended
  2. December 2025: not recommended
  3. January 2026: not recommended
  4. February 2026: not recommended
  5. March 2026: not recommended
  6. April 2026: not recommended
  7. May 2026: recommended
  8. June 2026: not recommended
  9. July 2026: recommended
  10. August 2026: recommended
  11. September 2026: recommended
  12. October 2026: not recommended
  13. November 2026: not recommended (month in progress)

Why it doesn’t carry the badge

  • 9 people recommend itNeeds 10 or more, not met.
  • 9 recommenders for every criticNeeds 2 to 1 or better, met.
  • Recommended in 5 of 13 monthsNeeds 4 or more, met.

The same three rules apply to every product, and the badge can’t be bought. How the badge works

What people say about it

The reasons people give, with the number of comments that give each one. Open a reason to read what they said.

Why people recommend it

Each mark is one comment

  • Minimal, simple, compact codebase

    “an open source agent that uses a minimal amount of CPU and RAM and has source that is easy to audit”

    Comment from 11 Aug 2026 in “Show HN: Ante, a coding agent in a single binary that runs offline”

    “smol is very minimal only using stdlib”

    Comment from 11 Aug 2026 in “Show HN: Ante, a coding agent in a single binary that runs offline”

    “Smol is simple and compact.”

    Comment from 21 Aug 2026 in “Malicious Rust Crate Arrayref Runs a Build-Time Payload”
  • Easy to read and adapt

    “if you want to see what a simple harness implementation looks like you can just dive into the code”

    Comment from 27 Aug 2026 in “Harness Engineering”

    “you can read each run step by step and see what the agents are doing”

    Comment from 4 Aug 2026 in “Agent skills that bring team coding standards to Claude Code and Codex”

    “because smol is so smol you can adapt it easily”

    Comment from 5 Aug 2026 in “Pi's Minimalism Is Its Advantage”
  • Alternative to Docker isolation

    “that's a great way to leverage hypervisor isolation to avoid those issues”

    Comment from 11 May 2026 in “Ask HN: What Are You Working On? (May 2026)”

    “definitely enjoying smol. this is a very nice alternative to docker”

    Comment from 3 Jul 2026 in “Podman v6.0.0”
  • Fits easily in context window

    “it fits into the context window with plenty of room to spare”

    Comment from 5 Aug 2026 in “Pi's Minimalism Is Its Advantage”

    “look at minimal agents that protect the context window”

    Comment from 8 Aug 2026 in “Databricks drove down AI coding spend 70%”

What people warn about

Each mark is one comment

  • Missing subagents and system prompt

    “smol currently is very simple so it definitely does less things, like no subagent orchestration”

    Comment from 1 Aug 2026 in “Cursor removed cost information from the usage page and CSV export”

    “in smol (minimalist agent harness) there currently is no system prompt at all”

    Comment from 16 Aug 2026 in “Claude: System Prompts”
  • Weak caching and token efficiency

    “so in this specific case the count of uncached tokens for smol makes it look worse than it actually is”

    Comment from 1 Aug 2026 in “Cursor removed cost information from the usage page and CSV export”

    “the token efficiency hurts cache use as caching only starts at 1024 tokens”

    Comment from 1 Aug 2026 in “Cursor removed cost information from the usage page and CSV export”

Other coding agent harnesses people recommend

The four most recommended other coding agent harnesses, with the same counts for each, so you can compare like for like. The tick on each bar marks the 2 to 1 line the badge needs.

smol compared with 4 alternatives
ProductRecommendCriticizeApprovalMonthsOwnedBadge
Pi coding agent89219882%123 mo51 ownersApproved
omp (oh-my-pi)802179%6—Approved
ForgeCode391276%10—Approved
Reasonix261958%6——
smolThis page9190%5——
See all 11 coding agent harnesses, ranked

What people said

22 of the 24 comments that mention it, newest first. Each links to the original comment.

“Smol was the only one to answer a basic question about javascript correctly”

RecommendsComment from 29 Sep 2026 in “MicroLLM Lab – Try 7 tiny LLM's in the browser”

“desktop applications or those who prefer the simplicity”

RecommendsComment from 30 Aug 2026 on Hacker News

“if you want to see what a simple harness implementation looks like you can just dive into the code”

RecommendsComment from 27 Aug 2026 in “Harness Engineering”

“Smol is simple and compact.”

RecommendsComment from 21 Aug 2026 in “Malicious Rust Crate Arrayref Runs a Build-Time Payload”

“in smol (minimalist agent harness) there currently is no system prompt at all”

MixedComment from 16 Aug 2026 in “Claude: System Prompts”

“smol is very minimal only using stdlib”

RecommendsComment from 11 Aug 2026 in “Show HN: Ante, a coding agent in a single binary that runs offline”

“an open source agent that uses a minimal amount of CPU and RAM and has source that is easy to audit”

RecommendsComment from 11 Aug 2026 in “Show HN: Ante, a coding agent in a single binary that runs offline”

“look at minimal agents that protect the context window”

RecommendsComment from 8 Aug 2026 in “Databricks drove down AI coding spend 70%”

“why is it codegolfed to such an extent?”

MixedComment from 6 Aug 2026 in “Prime Agent: A self-improving RLM agent”

“it makes a decent starting point”

RecommendsComment from 5 Aug 2026 in “Pi's Minimalism Is Its Advantage”

“sh is pretty versatile and composes well”

RecommendsComment from 5 Aug 2026 in “Pi's Minimalism Is Its Advantage”

“because smol is so smol you can adapt it easily”

MixedComment from 5 Aug 2026 in “Pi's Minimalism Is Its Advantage”

“it fits into the context window with plenty of room to spare”

RecommendsComment from 5 Aug 2026 in “Pi's Minimalism Is Its Advantage”

“I like the smol philosophy.”

RecommendsComment from 4 Aug 2026 in “Agent skills that bring team coding standards to Claude Code and Codex”

“you can read each run step by step and see what the agents are doing”

RecommendsComment from 4 Aug 2026 in “Agent skills that bring team coding standards to Claude Code and Codex”

“smol looks interesting i think it could have a potential niche although tokens are only going to get cheaper and cheaper here”

MixedComment from 2 Aug 2026 in “Cursor removed cost information from the usage page and CSV export”

“smol currently is very simple so it definitely does less things, like no subagent orchestration”

MixedComment from 1 Aug 2026 in “Cursor removed cost information from the usage page and CSV export”

“the token efficiency hurts cache use as caching only starts at 1024 tokens”

CriticizesComment from 1 Aug 2026 in “Cursor removed cost information from the usage page and CSV export”

“so in this specific case the count of uncached tokens for smol makes it look worse than it actually is”

RecommendsComment from 1 Aug 2026 in “Cursor removed cost information from the usage page and CSV export”

“definitely enjoying smol. this is a very nice alternative to docker”

RecommendsComment from 3 Jul 2026 in “Podman v6.0.0”

“that's a great way to leverage hypervisor isolation to avoid those issues”

RecommendsComment from 11 May 2026 in “Ask HN: What Are You Working On? (May 2026)”

“Smol is multiplicatively faster than B-Trees and radically smaller.”

RecommendsComment from 9 Nov 2025 in “Ask HN: What Are You Working On? (Nov 2025)”

Counts come from 1,795,640 Hacker News comments, November 2025 to November 2026. How we count. Report a mistake.