Skip to main content

issue-implement

Browse tracked issues and pick one to implement. Lists issues via bluectl, reads the implementation plan from the issue notes, and produces an actionable plan to start working on it.

Aller à l'installation

Informations de source

Dépôt
unstablebuild/blue
Dernière activité de la source
11 mars 2026 à 08:21
Langue détectée de SKILL.md
anglais
Étoiles
3
Forks
0

Options d'installation

Le prompt qui vérifie d'abord la source est sélectionné par défaut. Vous pouvez passer à une commande directe ou télécharger une copie locale.

Vérifiez les fichiers source

Lisez SKILL.md et les fichiers associés affichés par SkillsMP avant de décider de l'installer.

Affichage de SKILL.md

SKILL.md
Instructions source · Aperçu en lecture seule
name
issue-implement
description
Browse tracked issues and pick one to implement. Lists issues via bluectl, reads the implementation plan from the issue notes, and produces an actionable plan to start working on it.
allowed-tools
Bash(bluectl:*) Read Write Edit Glob Grep Agent
# Issue Implement Skill You help the user pick a tracked issue and turn it into work. ## Conversation flow ### Step 1 - List issues Ask the user which package to list issues for. First, list issues that have an implementation plan ready: ``` bluectl issue list <package> -f ready=true ``` Then list all remaining open issues: ``` bluectl issue list <package> ``` Present both lists to the user, clearly separating them: **Ready to implement** (have an implementation plan): <table from the ready=true filter> **All open issues:** <table from the unfiltered list> Ask which issue they want to work on. Recommend starting with a "ready" issue since those already have a plan. ### Step 2 - Fetch the issue Run `bluectl issue get <id>` with the chosen issue ID. Parse the output. The `notes` field may contain: - A plain description (older issues) - A structured plan with "Relevant code" and "Implementation plan" sections (issues created by `/issue-create`) ### Step 3 - Assess the plan **If the issue already has an implementation plan** in its notes: - Present the plan to the user. - Ask if they want to proceed as-is, adjust it, or do fresh research first. **If the issue has no implementation plan:** - Summarize the issue for the user. - Ask which repositories to look at (default: current working directory). - Spawn an Explore agent to research the relevant code, looking for files, types, and patterns related to the issue. - Produce an implementation plan based on the research. - Present it to the user for confirmation. ### Step 4 - Produce the final plan Once the user confirms, output the plan the implementing agent should follow: ``` ## Implement: <issue-id> - <subject> ### Context <brief summary of what the issue asks for> ### Implementation steps 1. <concrete step referencing specific files/symbols> 2. ... ### Verification - <how to verify each step works: tests to run, commands to check, etc.> ### Close the issue Once all steps are verified, run: `bluectl issue close <issue-id>` ``` The implementing agent (or the user) takes it from here. ## Rules - Be concise. Ask one question at a time. - If the user provides an issue ID directly (e.g. `/issue-implement PKG-4`), skip straight to Step 2. - When an issue already contains a plan, prefer using it over doing redundant research. Only re-research if the user asks or the plan references files that no longer exist. - Implementation steps must reference real files and symbols, not hypothetical ones. - Always include a verification section so the agent knows how to confirm the work is correct.
Voir sur GitHub