feat: a skill that finds where the session's skills didn't fit and files the gaps
I maintainer di solito rispondono entro 1 giorno
Nessuno ha ancora preso questa issue.
Valutazione
- Difficoltà
- 5/5
- Tempo stimato
- Più di una settimana
- Idoneità per principianti
- 22/100
- Tipo di issue
- Funzionalità
- Chiarezza
- Abbastanza chiara
- Stato di attività
- Attiva
- Stack tecnologico
- shell
- Ambito
- ai, developer-experience, tooling
Direzione di ricerca
This is a new skill (suggested names /plugin-quality:fit-gaps or /session-flow:skill-fit) that must read on-disk session transcripts and subagent returns, attribute gaps to owning plugins, de-duplicate with gh issue list --search, and only file issues after owner confirmation. Start from the existing skills named in the issue (/discipline:use-your-skills, /plugin-quality:audit, /session-flow:retro) and from session dba79e68-af19-4ecc-a0ce-69d8b1dbefbb plus issues #6412 and #6413 as the intended output. Done means the skill presents a ledger and files evidence-first issues only on yes; this is product design, not a one-file edit.
Scritto dal modello di indicizzazione a partire dal testo della issue.
Descrizione
Problem
When a skill doesn't fit the work in front of it, the session quietly works around it. That workaround is the most useful feedback a plugin can get, and today it is lost unless someone notices and files it by hand.
In the 2026-10-04 CI-performance session the owner asked for that by hand twice. Each time the coordinator reconstructed the gaps from memory and filed #6412 and #6413. The gaps it found:
- implement-dispatch needs a PLAN.md, but the work came as scoped audit output, so workers were hand-dispatched with re-typed lifecycle rules;
- workers stalled five times waiting on their own background work;
- a research skill's verifier failed twice on labeling for vendor-documented facts, so the loop was stopped by judgment;
- workers' classifier denials could not see the owner's approval, so edits were escalated to the main session.
Existing skills cover parts but not this:
/discipline:use-your-skillsfinds a skill that should have fired, not one that fired and didn't fit./plugin-quality:auditaudits one named component after use, not a whole session's fit./session-flow:retrocodifies session learnings, not per-skill gaps routed to their owners.
Ask
A skill the owner can fire at any point (start, middle or end), for example /plugin-quality:fit-gaps or /session-flow:skill-fit. It would:
- Read the session's on-disk transcript and its subagents' returns. Find each place where:
- a skill was invoked and then bypassed or overridden;
- a coordinator re-typed the same rules into several briefs;
- a worker stalled, or was nudged or re-briefed for the same reason more than once;
- a permission or classifier denial forced a workaround;
- a verifier loop was stopped by judgment;
- the owner asked "is there a skill for this".
At session start, with no transcript yet, it instead lists the skills that likely apply to the stated task and their known fit limits, from open issues.
- Attribute each gap to its owning plugin or skill, citing transcript evidence.
- De-duplicate against that plugin's open issues (
gh issue list --search). - Present a ledger of gap, owning skill, evidence, proposed change and existing issue. File the new ones only on the owner's yes, with evidence-first bodies. That fits the org's rule that no automated process opens issues.
- Optionally propose a new-skill candidate when no existing skill owns a recurring pattern.
Evidence
Session dba79e68-af19-4ecc-a0ce-69d8b1dbefbb, plus #6412 and #6413, which this skill would have produced.
- Lingua principale
- Shell
- Stelle
- 22
- Fork
- 2
- Merge medio
- 5h 15m
- PR unite (30g)
- 833
Preparare l'ambiente
- Nessun Dockerfile né file Docker Compose
- Nessun modello di pull request
- Leggi la guida per i contributori
Come iniziare
- Leggi tutta la issue e poi la guida ai contributi del progetto.
- Commenta sulla issue per dire che te ne occupi tu — evita che due persone facciano lo stesso lavoro.
- Fai un fork del repository e lavora su un branch.
- Apri una pull request che faccia riferimento al numero della issue.
Altre issue di melodic-software/claude-code-plugins
-
good first issue needs-triage priority: medium
Difficoltà 2/5 1-3 ore Idoneità per principianti 72/100
melodic-software/claude-code-plugins#6631 · 1 commento ·
I maintainer di solito rispondono entro 1 giorno
-
needs-triage
Difficoltà 2/5 1-3 ore Idoneità per principianti 78/100
melodic-software/claude-code-plugins#6547 ·
I maintainer di solito rispondono entro 1 giorno
-
needs-triage
Difficoltà 2/5 1-3 ore Idoneità per principianti 76/100
melodic-software/claude-code-plugins#6535 ·
I maintainer di solito rispondono entro 1 giorno
-
test_comment_census.py: SccArgv flag-shaped-filename test errors on Windows (#!/bin/sh scc shim)Apertagood first issue needs-triage priority: low
Difficoltà 2/5 1-3 ore Idoneità per principianti 72/100
melodic-software/claude-code-plugins#6532 · 1 commento ·
I maintainer di solito rispondono entro 1 giorno
-
good first issue needs-triage priority: low
Difficoltà 2/5 1-3 ore Idoneità per principianti 72/100
melodic-software/claude-code-plugins#6390 · 1 commento ·
I maintainer di solito rispondono entro 1 giorno
Tutte le issue di melodic-software/claude-code-plugins
Issue simili
-
Difficoltà 2/5 1-3 ore Idoneità per principianti 82/100
aws-samples/appmod-blueprints#972 ·
I maintainer di solito rispondono entro 1 giorno
-
status:needs-triage
Difficoltà 2/5 1-3 ore Idoneità per principianti 72/100
I maintainer di solito rispondono entro 1 giorno
-
Difficoltà 2/5 1-3 ore Idoneità per principianti 62/100
I maintainer di solito rispondono entro 1 giorno
-
go: new version 1.27.2Apertaout-of-date
Difficoltà 2/5 1-3 ore Idoneità per principianti 62/100
CachyOS/CachyOS-PKGBUILDS#1965 ·
I maintainer di solito rispondono entro 1 giorno
-
enhancement good first issue help wanted
Difficoltà 2/5 1-3 ore Idoneità per principianti 82/100
Sorogate/example-consumer#14 ·