skill-feedback
Skill1 starsChanged 4 months agoMITView on GitHub
Log a problem or bad output from any skill so it can be improved later. Use when the user says a skill produced wrong output, missed something, was unclear, or triggered when it should not have. Also triggers on "log feedback", "report a skill issue", "something went wrong with the skill".
7 other projects use a version of this.
Adopted by
Most recently taken up by
- adeonir/agent-skillstheir own version, first seen Nov 1, 2026
- ingoclaro/claude-skillstheir own version, first seen Oct 20, 2026
- microsoft/microsoft-sqltheir own version, first seen Oct 13, 2026
- caikiji/skills-labtheir own version, first seen Sep 7, 2026
- bestdeejay-design/agent-skillstheir own version, first seen Aug 9, 2026
- bstegemoller/skill-feedbacktheir own version, first seen Jul 10, 2026
Other versions7 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 7
- microsoft/microsoft-sql★ 22same name
- adeonir/agent-skills★ 8same name
- bestdeejay-design/agent-skills★ 6same name
- caikiji/skills-lab★ 4same name
- adamkrasny/adams-claude-skills★ 1same name
- bstegemoller/skill-feedback★ 0same name
- ingoclaro/claude-skills★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository2 other files it gives its agents.
- commit-messageSkill
- skill-improverSkill
The file
We split this skill into the rules it gives an agent
8 rules. 0 are common, 4 are distinctive, and 4 are very specific to this project.
0
Common
Most of the 6 edited versions we read keep these.
4
Distinctive
This author's own choices. Worth a look.
4
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 20
Do not accept meta-skills (skill-feedback, skill-improver) as targets.
- Distinctive: line 21
If the user names one of those, say: "Meta-skills cannot be logged as feedback targets.
- Distinctive: line 26
Ask for any inputs not already present in the user's message.
- Distinctive: line 27
Ask all missing inputs in a single message, never one at a time.
- Very specific: line 40
Read ~/.claude/skill-data/feedback-log.md to confirm it exists.
- Very specific: line 63
Run $skill-improver when you want to batch-process feedback."
- Very specific: line 67
Never edit any SKILL.md file.
- Distinctive: line 68
Never modify existing entries in feedback-log.md.
“Common” is measured: kept by at least 6 in 10 of the 6 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Skill Feedback
Purpose
Collect structured feedback about any domain skill and append one entry to
~/.claude/skill-data/feedback-log.md. Do not modify any skill file. Do not
attempt to fix anything. Only log.
Step 1 — Identify the skill
The user must name the skill. It must match a folder name inside
~/.claude/skills/. If it does not match, list the available folders and ask
the user to confirm which one they mean.
Do not accept meta-skills (skill-feedback, skill-improver) as targets.
If the user names one of those, say: "Meta-skills cannot be logged as feedback
targets. Did you mean one of your domain skills?" and list them.
Step 2 — Collect required inputs
Ask for any inputs not already present in the user's message. Ask all missing inputs in a single message, never one at a time.
Required:
- section — which part of the output was wrong, described as concisely as possible (e.g. "test steps", "acceptance criteria", "trigger", "format").
- issue — what exactly was wrong. One or two sentences.
- severity — integer 1 to 5 (see scale in feedback-log.md).
Optional:
- context — the original prompt that caused the issue.
Step 3 — Append the entry
Read ~/.claude/skill-data/feedback-log.md to confirm it exists. If it does
not exist, stop and say: "feedback-log.md not found at
~/.claude/skill-data/feedback-log.md — run install.sh from the
claude-skill-improver repo to create it."
Append a new block at the very bottom of the file, after the line
<!-- $skill-feedback appends below this line -->, using this exact format
with no extra whitespace or blank lines between fields:
---entry---
skill: <skill>
date: <today YYYY-MM-DD>
severity: <severity>
section: <section>
issue: <issue>
context: <context or "none">
status: open
---entry---
Step 4 — Confirm
Reply with exactly:
"Logged. Feedback added for <skill> — section: <section>, severity: <severity>/5.
Run $skill-improver when you want to batch-process feedback."
Rules
- Never edit any SKILL.md file.
- Never modify existing entries in feedback-log.md. Only append.
- One entry per invocation. If the user mentions multiple issues, log the most severe one and ask them to invoke $skill-feedback again for the others.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.