lfnovo/harny/learn
Claude Code and agent skill by lfnovo/harny
- community
Checking that the SKILL.md file is still in the repository.
Append a one-line learning to the local harny inbox. Fast, non-analytical capture — copies text verbatim, no analysis, no follow-ups. Use when the user wants to note something about harny adoption mid-conversation without breaking flow.
Install
npx vskill@latest install lfnovo/harny/learnSKILL.md
learn — capture a harny learning
Fast, frictionless capture into the project's local harny inbox at <cwd>/.harny/LEARNINGS.local.md. Drained later by the drain skill.
The whole point is zero friction. The user is mid-conversation and wants to note something without breaking flow. Do not analyze. Do not ask follow-ups. Do not propose solutions. Capture and return.
Steps
1. Locate the inbox
Path: <cwd>/.harny/LEARNINGS.local.md. The cwd is the current working directory of the conversation.
If .harny/ does not exist, create it. The file .harny/.gitignore (containing * + !.gitignore) ships with harny runs and keeps everything inside .harny/ out of git, including this inbox. If .harny/ is being created here for the first time, also write that gitignore.
2. Lazy-create the inbox file if missing
Use this header exactly:
3. Gather minimal context
- Current ISO UTC timestamp. Example:
2026-04-23T18:45:00Z. Usedate -u +%Y-%m-%dT%H:%M:%SZvia Bash. - Current git branch via
git rev-parse --abbrev-ref HEAD. If the repo is detached orgiterrors, use-.
4. Append the entry
Single-line format:
If the captured text has newlines, indent continuation lines with two spaces so the bullet list stays valid:
Copy the user's text verbatim. Do not rewrite, do not summarize, do not "improve" the wording.
5. Confirm briefly
One short line back to the user, then return to whatever was being discussed:
Captured to.harny/LEARNINGS.local.md. Back to what we were doing.
No questions. No analysis. No "would you like me to also...".
What this skill does NOT do
- Does not apply the counterfactual test ("would a fresh dev hit this?").
- Does not propose file targets (AGENTS.md edits, GitHub issues, etc.).
- Does not open issues or discussions.
- Does not edit any file other than
.harny/LEARNINGS.local.mdand (on first creation).harny/.gitignore. - Does not prompt for more detail.
All of that belongs to the drain skill, invoked separately when the user is ready to triage accumulated entries.
Edge cases
- No `args` provided. Ask one short question: "What do you want me to capture?" — then return to the steps above. Do not ask for elaboration on the content.
- `args` starts with the word `drain`. Do not run capture. Tell the user this skill is for capture only and suggest
/drainfor triage. Do not invokedrainautomatically. - `.harny/` exists but is not a directory (e.g., a file with that name). Stop and report — do not overwrite.
- `git` command unavailable. Use
-for the branch field rather than failing.
Security
- Security scanPassed, 90 of 100Pattern checks on the skill's files.
- Model reviewPassed
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingplugin/skills/learn/SKILL.md
Popularity
- Repository stars
- 11
- Repository forks
- 5
- 7-day trend score
- 0
Stars and forks are for the whole GitHub repository, not this skill alone. Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/lfnovo/harny/learn)