Sigma lessons

Skill navidgh66/sigma/skills/sigma-lessons

Recall sigma's past ratcheted lessons before implementing or verifying a task, so previous mistakes are not repeated. Use when about to write or review code in any sigma domain (classic-ml, deep-learning, nlp, rl, data-analysis, data-engineering, ai-agent-engineering, mlops, llm-engineering), when starting a loop task, or when the user asks "what have we learned" / "avoid past mistakes". Lessons are written by the loop on failure and by /sigma-learn-lesson; this skill reads them back.From its SKILL.md

Install

npx -y skills add navidgh66/sigma --skill sigma-lessons

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

3 stars3 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

2.0 KB, as published. Nobody here has run it

sigma-lessons

sigma accumulates lessons from failures (the loop ratchets them) and from the human (/sigma-learn-lesson). Each lesson is a skills/<slug>/SKILL.md tagged with metadata: domain:. This skill reads them back so a future task applies them instead of repeating the mistake.

When to use

Before implementing or verifying a task in a known domain — pull the lessons for that domain first.
When the user asks to recall what was learned, or to avoid prior mistakes.

Workflow

Identify the domain of the current task (one of the 9 sigma domains; if a task line is annotated (domain), use that).
Read the lessons for that domain: scan skills/**/SKILL.md for files whose frontmatter domain: matches. Each has a **Lesson (ratcheted):** line and a **How to apply:** line — those are the actionable parts.
Apply them as constraints while implementing or as extra checks while verifying. Treat a lesson as "do not repeat this mistake."
If a lesson looks stale or wrong, do NOT silently delete it — flag it for the human (sigma never auto-resolves lessons; see skills/CONTRADICTIONS.md).

Notes

The CLI loop does this automatically (cli/skills_recall.py builds the recall block and injects it into the implement + verify prompts). This skill is the in-session equivalent for slash-command work.
Lessons without a domain: (vendor skills, sigma-present, sigma-domains) are NOT lessons — ignore them here.