j4flmao/agent-skills/agent-task-checklist
Claude Code and agent skill by j4flmao/agent-skills
- community
Checking that the SKILL.md file is still in the repository.
Use this skill when the user says 'track progress', 'task.md', 'keep a checklist', 'break this into steps', 'show me what is done', or when an agent executes a long multi-step task that should remain transparent and resumable. It maintains a granular, frequently updated checklist with explicit states ([ ], [/], [x]) so progress survives interruptions and the user can see status at a glance. Do NOT use for: the up-front design of changes (use write-implementation-plan), sprint/backlog management for human teams (use agile-scrum), final reporting (use agent-walkthrough-report).
Install
npx vskill@latest install j4flmao/agent-skills/agent-task-checklistSKILL.md
Agent Task Checklist
Purpose
Make long agent tasks observable and resumable. A checklist file (task.md, or the host's built-in task-list tool when one exists) lists atomic steps and their state, and is updated the moment a step starts or finishes — so a user, a reviewer or a resumed session can tell exactly what is done, in progress and left.
Agent Protocol
Trigger
"task.md", "checklist", "track progress", "break it down", "what's left", long tasks (>5 steps or multiple sessions).
Input Context
- The approved plan or the user's request.
- Whether the host provides a native task/plan tool (prefer it over a file).
Output Artifact
task.md at the repo root or scratch location agreed with the user (never committed unless asked).
Response Format
State Notation
| Mark | Meaning | Rule | |
|---|---|---|---|
[ ] | Not started | Default for new items | |
[/] | In progress | Exactly one item at a time unless work is truly parallel | |
[x] | Done | Only after verification (tests ran, file saved, command succeeded) | |
[-] | Dropped | Keep with a reason instead of deleting silently |
[/] is a convention of this workflow, not standard Markdown task-list syntax; GitHub renders only [ ] and [x] as checkboxes.
Granularity Rules
- One step = one verifiable outcome, typically minutes, not hours.
- Start each step with a verb and name the file or command: "Add X to
path", "Runcmd". - Split any step that touches more than one component or has its own verification.
- Put verification as its own step ("Run tests", "Check build") so it cannot be skipped implicitly.
Update Discipline
- Write the full list before starting work.
- Mark
[/]when beginning a step;[x]immediately when it is verified. - When new work is discovered, insert it in order — do not silently expand an existing step.
- When the plan changes, rewrite the list so it never shows abandoned steps as pending.
- On interruption or handoff, the checklist plus Notes must be enough to resume.
Anti-Patterns
| Anti-pattern | Fix | |
|---|---|---|
| Ticking everything at the end | Update per step | |
| Vague items ("fix stuff") | Verb + path/command | |
| Marking done without running checks | Verify, then tick | |
Several [/] at once | One active item | |
| Checklist and reality diverge after a plan change | Rewrite the list |
Checklist
- [ ] All steps written before execution
- [ ] Exactly one in-progress item
- [ ] Verification steps explicit
- [ ] Blockers recorded with owner
- [ ] Final state shows no stale
[ ]items
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 repositoryCheckingskills/planning/agent-task-checklist/SKILL.md
Popularity
- Repository stars
- 27
- Repository forks
- 1
- 7-day trend score
- 22
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/j4flmao/agent-skills/agent-task-checklist)