Home / robium-ai / robium · skills/learning-loop/SKILL.md · GitHub

learning-loop skillA

learning-loop is agent-read markdown (skill) from robium-ai/robium: Turn captured Robium experience into small, evidence-backed skill improvements..

Indexed from public GitHub and served as immutable, content-addressed versions. Install it pinned to an exact SHA-256 with the mdr CLI, and every file is verified against the hash recorded here before it reaches your agent. The deterministic audit below grades the latest version, and the same file always earns the same grade.

What the file says

# Learning loop

Capture is a lead, not knowledge. Promote it only when the evidence can improve
a future decision without injecting old conversations into new work.

## Capture without recall

- Hooks may queue corrections, failures, and transcript windows silently.
- Findings live in one store, `learnings/observations/<skill>.md`. Capture is
  a `tentative` entry there — status and signal are the only required fields.
- Never insert queue items, observations, memories, or reminders into a new
  prompt. Read them only during an explicit consolidate, absorb, refine, or
  status task.
- Batch bookkeeping at a natural milestone. A command error or clean skill run
  is not automatically a reusable lesson.
- Keep project-specific facts in the application that owns them.

## Promote evidence

- Read the source transcript before turning a flag into an observation.
- Preserve what was expected, what actually happened, the evidence for the
  fix, and any dead ends that prevent repeating the mistake.
- Absorb only observations marked `ready`; uncertain material stays tentative.
- Deduplicate the finding before carrying it forward.
…

Read the whole file at its exact version.

How to install

Latest version
mdr add robium-ai/robium/learning-loop@git:20260922.d17aa3e
Exact content
mdr add robium-ai/robium/learning-loop@sha256:3eef4b9c680b5b3b

Pin to a label to follow the author's releases, or to a sha256 to freeze the exact bytes forever. Either way the resolved hash is written to mdr.lock, and mdr install reproduces it on any machine.

Badge

mdr badge

[![mdr](https://markdownregistry.com/badge/art_3agqbcmqa5y2r2ho.svg)](https://markdownregistry.com/a/art_3agqbcmqa5y2r2ho)

1 badge views in 30 days

Versions

versioncommittedcommitsizeaudit
git:20260922.d17aa3e latest2026-09-22 d17aa3e 3,275 BA view · diff
git:20260907.863840e2026-09-07 863840e 3,061 BA view · diff
v1.2.02026-08-28 371219b 11,293 BA view · diff
v1.1.02026-08-28 a7a0334 11,151 BA view · diff
v1.0.12026-08-28 32770d5 10,676 BA view

Audit of the latest version

A  17 of 17 checks passed. Deterministic, no model, same answer every run.
  • pass: Frontmatter block present
  • pass: Frontmatter declares a name
  • pass: Frontmatter declares a description
  • pass: Size between 200 bytes and 200 KB (3275 bytes)
  • pass: No zero-width or bidi control characters
  • pass: No instruction hidden inside an HTML comment
  • pass: No link to an exfiltration or paste host
  • pass: No credential-shaped string
  • pass: No instruction to send local credentials anywhere
  • pass: No text hidden with inline styles
  • pass: No prompt-injection phrasing
  • pass: No curl or wget piped into a shell
  • pass: No recursive delete of root, home or parent
  • pass: No instruction to read or print local credentials
  • pass: No base64 blob over 200 characters
  • pass: No link to a raw IP address
  • pass: No script tag

Source

GitHub

robium-ai/robium · 14 stars · license MIT · pushed 2026-09-22 · branch main

API

GET https://markdownregistry.com/api/v1/artifacts/art_3agqbcmqa5y2r2ho
GET https://markdownregistry.com/api/v1/resolve?ref=robium-ai/robium/learning-loop
GET https://markdownregistry.com/api/v1/blob/3eef4b9c680b5b3b13b787219d04496016e4aa918d756254697c71d22c7c177a

Your agent does the legwork. You hear about the deals worth your word. Hand yours the standing instructions at modelranch.com and it joins the network that reads files like this one.

More from robium-ai/robium

AGENTS.md agents
robium-ai/robium · AGENTS.md
git:20260907.863840e · audit A · 14 stars
CLAUDE.md claude
robium-ai/robium · CLAUDE.md
git:20260807.b408a93 · audit A · 14 stars
alpha skill
robium-ai/robium · cli/test/fixtures/skills/alpha/SKILL.md · A folded-scalar description that spans multiple indented lines and should be joined with single spaces. Second sentence…
git:20260907.863840e · audit A · 14 stars
beta skill
robium-ai/robium · cli/test/fixtures/skills/beta/SKILL.md · Plain single-line description.
git:20260907.863840e · audit B · 14 stars
AGENTS.md@learnings agents
robium-ai/robium · learnings/AGENTS.md
git:20260922.d17aa3e · audit A · 14 stars
CLAUDE.md@learnings claude
robium-ai/robium · learnings/CLAUDE.md
git:20260922.9c918e1 · audit B · 14 stars
AGENTS.md@skills agents
robium-ai/robium · skills/AGENTS.md
git:20260922.c5a4f3f · audit A · 14 stars
CLAUDE.md@skills claude
robium-ai/robium · skills/CLAUDE.md
git:20260922.9c918e1 · audit B · 14 stars
app-publishing skill
robium-ai/robium · skills/app-publishing/SKILL.md · Publish a tested Robium application as a coherent project page, article, and media set.
git:20260907.863840e · audit A · 14 stars
architect skill
robium-ai/robium · skills/architect/SKILL.md · Start robotics apps or demos from compatible examples for mapping/navigation, robot assistants, or pretrained…
git:20260919.498ea4e · audit A · 14 stars
cloud-run skill
robium-ai/robium · skills/cloud-run/SKILL.md · Deploy and debug headless robotics containers on Google Cloud Run.
git:20260907.863840e · audit A · 14 stars
data skill
robium-ai/robium · skills/data/SKILL.md · Choose and structure training data for robot-learning projects.
git:20260918.06c6019 · audit A · 14 stars

Every file in robium-ai/robium

Other files named learning-loop

learning-loop skill
richfrem/agent-plugins-skills · plugins/agent-orchestration/skills/learning-loop/SKILL.md · (Industry standard: Loop Agent / Single Agent) Primary Use Case: Self-contained research, content generation, and…
git:20260915.2b5498a · audit A · 6 stars
learning-loop skill
wan-huiyan/agent-traffic-control · plugins/agent-traffic-control/skills/learning-loop/SKILL.md · Turn failures and user corrections into tested workflow improvements. Use for lessons and retrospectives.
git:20260912.c635a3a · audit A · 3 stars

Browse by kind, by grade A, or by owner.