Agent skill

Agentic Learn

by carlos-algms in carlos-algms/agentic.nvim

Turns rule gaps into routed, verified rules and empties rules-report.md.

MITAuto-check passedAgent Workflows

Install Agentic Learn

skills CLI
$ npx skills add carlos-algms/agentic.nvim --skill agentic-learn -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install carlos-algms/agentic.nvim agentic-learn --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/carlos-algms/agentic.nvim.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.agents/skills/agentic-learn .claude/skills/agentic-learn && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
agentic-learn
GitHub stars
642
Token cost
~796 tokens
SKILL.md length
417 words
Files
1
Skills in repo
8
Repo updated
First seen
Licence
MIT

At a glance

Turns rule gaps into routed, verified rules and empties rules-report.md.

  • Works in 5 steps: Verify it against the current code. Read… → wrong: fix or delete the rule. The code… → missing: write the rule where the router… → …
  • A rule was missing
  • SKILL.md covers Triggers, rules-report.md, Processing a gap and CodeRabbit learnings export
  • Calls git and make

What it does

Agentic Learn is an agent skill from carlos-algms/agentic.nvim. Turns rule gaps into routed, verified rules and empties rules-report.md. Use when a rule was missing, wrong, or broken; a review finding was fixed or rejected; a sub-agent report has a "Rule gaps" list; or rules-report.md exists.

Its SKILL.md is about 800 tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts.

It sits in Agent Workflows, covering Subagents and Code review. It works with Lua. The repository describes itself as: Agentic Chat Interface directly in Neovim with ACP providers from Claude-Code, Gemini, Codex, OpenCode, and Cursor-agent. The licence is MIT.

When your agent uses it

  • A rule was missing
  • A review finding was fixed
  • A sub-agent report has a Rule gaps list
  • Rules-report.md exists

Example prompts

  • “Rule gaps”
  • “Use the agentic-learn skill to turn rule gaps into routed, verified rules and empties rules-report.md”
  • “/agentic-learn”

Workflow steps

5 steps, taken from the first numbered list in SKILL.md.

  1. Verify it against the current code. Read the code; do not trust the report
  2. wrong: fix or delete the rule. The code is the truth unless the gap shows
  3. missing: write the rule where the router sends that file. Follow skill
  4. broken: the rule exists and was not read or not followed. Make it easier
  5. Remove the entry from rules-report.md. Delete the file when it is empty

What it can do on your machine

Read from SKILL.md and the folder at commit 5ee5110. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Shell commands in SKILL.md call:

    • git
    • make

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md. Its commands use git, which can reach the network depending on how they are called.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Agentic Learn loads about 796 tokens when it runs. Until then it costs about 61 tokens; SKILL.md has 417 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~61
When it runs · the whole SKILL.md, loaded when a task matches
~796

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from carlos-algms/agentic.nvim at commit 5ee5110, republished under its MIT licence (© carlos-algms). 417 words, ~796 tokens.

Download SKILL.mdSave it as .claude/skills/agentic-learn/SKILL.md (or your agent's skills folder).
name
agentic-learn
description
Turns rule gaps into routed, verified rules and empties rules-report.md. Use when a rule was missing, wrong, or broken; a review finding was fixed or rejected; a sub-agent report has a "Rule gaps" list; or rules-report.md exists.

Learn from rule gaps

The repo improves itself by writing down what agents trip on. A rule written next to the code it governs is read on the next task; a rule kept in someone's memory is not.

Triggers

What happenedQuestion to ask
You fixed a valid review findingWould a written rule have prevented it?
You rejected a review findingIs the rejection a fact about the whole codebase?
The user corrected you mid-taskWould a written rule have prevented it?
A rule did not match the codeWhich one is wrong: the rule or the code?
A sub-agent report has a Rule gaps listAppend it to rules-report.md, then process it
rules-report.md existsProcess every entry

A one-off suppression ("ignore this line") is not a rule. Skip it.

rules-report.md

The queue of gaps not fixed yet. It lives at the repo root, untracked and NOT gitignored, so git status shows it in every session until it is empty. Never stage or commit it; CI fails a PR that contains it.

Only the agent that receives sub-agent reports appends to it. Sub-agents never write it. One entry per gap:

markdown
- 2026-10-04 | lua/agentic/ui/chat_widget.lua | rule: ui/AGENTS.md "Traps" |
  gap: wrong | rule says X, code does Y

gap is one of missing, wrong, broken (the rule exists and was not followed).

Show full SKILL.md (212 more words)Show less

Processing a gap

  1. Verify it against the current code. Read the code; do not trust the report
  2. wrong: fix or delete the rule. The code is the truth unless the gap shows a bug; then fix the bug with TDD and keep the rule
  3. missing: write the rule where the router sends that file. Follow skill agentic-docs-authoring: a runtime rule cites a test that fails without it, and a banned token goes into the BANNED list in scripts/check-rules.sh, unless Selene or LuaLS already fails make validate on it (see skill agentic-docs-authoring)
  4. broken: the rule exists and was not read or not followed. Make it easier to find: move it closer to the code, or into the router, or make it mechanical (a check that fails). Rewording alone rarely helps
  5. Remove the entry from rules-report.md. Delete the file when it is empty

Rules go into the same change as the fix that revealed them. The human reviews them in the diff.

CodeRabbit learnings export

CodeRabbit keeps its own learnings. To move them into the repo, export learnings.csv from the CodeRabbit dashboard, sort by usage, and treat each prescriptive learning or API fact above about 100 uses as a missing gap. Leave suppressions ("intentional", "do not flag") in CodeRabbit.

© carlos-algms, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

Just SKILL.md in .agents/skills/agentic-learn of carlos-algms/agentic.nvim.

Open the folder on GitHubat commit 5ee5110

Compare with similar skills

Agentic Learn next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Agentic Learn compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Agentic Learn this skillcarlos-algms/agentic.nvim642—~796Automated safety check: PassMIT
O2 Review Loopopenobserve/openobserve22k—~3.7kAutomated safety check: PassAGPL-3.0
Elvesaigorahub/elves221—~16kAutomated safety check: NotesMIT
Subagent-Driven DevelopmentHoangNguyen0403/agent-skills-standard572—~1.3kAutomated safety check: PassMIT
Reviewjuliepy/AI-Engineer-from-scrach441—~417Automated safety check: PassNone
Simplifytellahq/opensession394—~834Automated safety check: PassMIT

Similar skills

  • O2 Review Loop

    openobserve/openobserve

    Splits a change into planner, coder and independent reviewer roles: you confirm a spec, a subagent implements it, and a separate reviewer checks each round's local WIP commit.

    22k GitHub stars~3.7k tokensUpdated yesterday
    Agent WorkflowsAuto-check passed
  • Elves

    aigorahub/elves

    Autonomous multi-batch development agent for long unattended runs, reviewed-PR landing, Cobbler-first orchestration, and optional Fugu, Manus, Grok, Devin, or Oh My Pi (omp) provider shortcuts.

    221 GitHub stars~16k tokensUpdated today
    Agent WorkflowsAuto-check: notes
  • Subagent-Driven Development

    HoangNguyen0403/agent-skills-standard

    Runs a multi-task implementation plan by sending each task to a fresh implementer subagent, reviewing it independently, then reviewing the whole branch.

    572 GitHub stars~1.3k tokensUpdated yesterday
    Agent WorkflowsAuto-check passed
  • Review

    juliepy/AI-Engineer-from-scrach

    Delegate the latest implementation diff to the code-reviewer sub-agent for a PASS/CONCERNS verdict against CLAUDE.md rules, then write the verdict to reports/<feature-review.md.

    441 GitHub stars~417 tokensUpdated 1 mo ago
    Agent WorkflowsAuto-check passed
  • Simplify

    tellahq/opensession

    Clean up the changed code without changing behavior — review the diff for reuse, simplification, efficiency, and altitude cleanups via 4 parallel agents, then apply the fixes

    394 GitHub stars~834 tokensUpdated yesterday
    Agent WorkflowsAuto-check passed
  • Agent Architect

    IgorWarzocha/Opencode-Workflows

    Create and refine OpenCode agents via guided Q&A. An agent skill from IgorWarzocha/Opencode-Workflows.

    122 GitHub stars~2.8k tokensUpdated 8 mo ago
    Agent WorkflowsAuto-check passed

More from carlos-algms/agentic.nvim

All 11 skills in this repo
  • Agentic GitHub Discussion

    carlos-algms/agentic.nvim

    Creates, lists, and manages GitHub discussions in this repo through the GraphQL API.

    642 GitHub stars~1.1k tokensUpdated yesterday
    Auto-check passed
  • Agentic Testing

    carlos-algms/agentic.nvim

    MANDATORY before creating, editing, or reviewing tests in agentic.nvim, and before behavior changes that require TDD.

    641 GitHub stars~828 tokensUpdated yesterday
    Auto-check passed
  • Agentic Lua Class

    carlos-algms/agentic.nvim

    Enforced Lua style: class pattern, visibility prefixes (private, protected), optional-type syntax, LuaCATS annotations.

    642 GitHub stars~1.6k tokensUpdated yesterday
    Auto-check passed
  • Agentic Vimdoc

    carlos-algms/agentic.nvim

    A skill your agent uses when writing or updating agentic.nvim vimdoc (doc/agentic.txt), or after changing init.lua, configdefault.lua, theme.lua, or README install/keymaps - those edits require a…

    641 GitHub stars~298 tokensUpdated yesterday
    Auto-check passed
  • Agentic Acp Docs And Schema

    carlos-algms/agentic.nvim

    Official ACP schema, events, and rules for this plugin's provider work.

    642 GitHub stars~481 tokensUpdated yesterday
    Auto-check passed
  • Agentic Self Review

    carlos-algms/agentic.nvim

    Reviews a diff before commit: tests can fail, routed rules hold for changed files, unchanged siblings still hold.

    642 GitHub stars~760 tokensUpdated yesterday
    Auto-check passed

Works with

Questions about Agentic Learn

What does Agentic Learn do?

Turns rule gaps into routed, verified rules and empties rules-report.md. nvim.md.

When should I use Agentic Learn?

Agentic Learn fits situations like: A rule was missing; A review finding was fixed; A sub-agent report has a Rule gaps list; rules-report.md exists.

How do I install Agentic Learn in Claude Code?

Run `npx skills add carlos-algms/agentic.nvim --skill agentic-learn -a claude-code`. Or copy the skill folder (.agents/skills/agentic-learn in carlos-algms/agentic.nvim) into .claude/skills/agentic-learn in your project. Claude Code loads it when a task matches its description.

How do I install Agentic Learn in Codex?

Run `npx skills add carlos-algms/agentic.nvim --skill agentic-learn -a codex`. Or copy the skill folder (.agents/skills/agentic-learn in carlos-algms/agentic.nvim) into .agents/skills/agentic-learn in your project. Codex loads it when a task matches its description.

Can I use Agentic Learn in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add carlos-algms/agentic.nvim --skill agentic-learn -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/agentic-learn, .gemini/skills/agentic-learn, .github/skills/agentic-learn and .opencode/skills/agentic-learn in your project.

What does Agentic Learn need to run?

Going by SKILL.md and its folder, Agentic Learn needs the command-line tools its instructions call (git and make).

Does Agentic Learn access the network?

SKILL.md contains no URLs. Its commands use git, which can reach the network depending on how they are called. This is read from the text; nothing was executed.

Is Agentic Learn safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Agentic Learn use?

Agentic Learn is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Agentic Learn use?

About 796 tokens (SKILL.md is roughly 3.2k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Agentic Learn?

Skills that share tags, products or a category with Agentic Learn: O2 Review Loop (openobserve/openobserve, 22k stars), Elves (aigorahub/elves, 221 stars), Subagent-Driven Development (HoangNguyen0403/agent-skills-standard, 572 stars) and Review (juliepy/AI-Engineer-from-scrach, 441 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Agentic Learn?

carlos-algms (a GitHub user) maintains it in carlos-algms/agentic.nvim, which has 642 GitHub stars. The repository holds 8 skills in this directory. The repository was last updated on October 10, 2026.

Source: carlos-algms/agentic.nvim on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.