agentleFS
lacerbilacerbi/dotclaude/skills/task/SKILL.md

task

Skill2 starsChanged 2 months agoMITView on GitHub

Create and maintain a task checklist that tracks progress in real-time

28 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use inbox, grill-me and 2 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • kv0906Skill

    inbox

    Quick capture mode (with content) or batch processing mode (no content). GTD-style routing of inbox items to proper folders. Use "/inbox thought to capture" or "/inbox" to process.

    Also in 3 projects that use this

  • GoogleSkill

    grill-me

    A relentless interview to sharpen a plan or design.

    Also in 5 projects that use this

  • VercelSkill

    setup

    Set up Vercel CLI and project configuration. Use when the user says "set up Vercel", "configure Vercel", "link to Vercel", or "vercel init".

    Also in 4 projects that use this

  • SentrySkill

    review

    Fixture review skill.

    Also in 3 projects that use this

The file

We split this skill into the rules it gives an agent

8 rules. 0 are common, 5 are distinctive, and 3 are very specific to this project.

  • 0

    Common

    Most of the 24 edited versions we read keep these.

  • 5

    Distinctive

    This author's own choices. Worth a look.

  • 3

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Keep it brief: one concise line per item, links to files instead of explanations (config).

    line 24
  2. Distinctive:

    When finished, the file shows the final status of all items, any unresolved issues (with explanation if incomplete), and a brief completion …

    line 40
  3. Very specific:

    Update the checklist to [~] before dispatching

    line 48
  4. Very specific:

    After completing all checklist items, run /doublecheck before marking the task complete.

    line 53
  5. Distinctive:

    Delete the TASK file if it contains only transient execution tracking, routine verification, and a completion summary.

    line 61
  6. Distinctive:

    If it records durable findings, decisions, constraints, or follow-ups, propose an appropriate destination based on repository conventions.

    line 62
  7. Distinctive:

    After approval, preserve the information there and delete the TASK file, unless archiving the file itself is appropriate.

    line 63
  8. Distinctive:

    Keep the file while work is incomplete or blocked.

    line 64

“Common” is measured: kept by at least 6 in 10 of the 24 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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Task

$ARGUMENTS

Your Instructions

1. Use Existing Plan or Create Task Plan

If a PLAN file was provided or mentioned (e.g., from a prior /plan command) with structured phases, it is your working document — do not create a separate TASK file.

Never replace or condense a PLAN file. It is the user's reviewed work product: goals, file references, verification details, rationale. Add tracking to it in place — convert implementation bullets to checkboxes (- Item → - [ ] Item), append new items or sections as gaps appear, update status markers as you progress — so that the checkboxes sit alongside the existing prose, never instead of it. An edit that removes more text than it adds is almost certainly wrong.

If no suitable PLAN file exists, create a markdown file TASK_[descriptor]_[timestamp].md containing:

  • Task description
  • Checklist of specific steps/subtasks
  • Success criteria
  • Any dependencies or prerequisites

Keep it brief: one concise line per item, links to files instead of explanations ([config](./src/config.js)). This is a tracking tool, not documentation.

Status markers:

  • [ ] Not started
  • [~] In progress
  • [x] Complete
  • [!] Blocked or needs attention

2. Execute While Updating

Update the checklist after every significant action — [~] when starting an item, [x] when done, [!] on blockers — and add subtasks as they surface. The file is the live record: sessions crash mid-run, and anyone (including a fresh session) must be able to read current progress from the file alone. Keep notes extremely brief or link to files.

3. Completion Expectations

Complete the full task/plan. Do not skip items, defer work, or stop halfway; minor deviations and unexpected issues are normal — handle them and continue. Stop early only for an overwhelming reason (core assumptions fundamentally wrong, blocking dependencies unresolvable).

When finished, the file shows the final status of all items, any unresolved issues (with explanation if incomplete), and a brief completion summary.

4. Sub-Agent Usage

Use the executor the plan names for each phase (see Agent Selection in the global CLAUDE.md). Carry out orchestrator phases yourself, whatever model the plan names for them; when it differs from yours, say so, since the user may want to switch models. Delegate the other phases to the named model as the global CLAUDE.md describes: Claude models through the Agent tool, Astra, Sol and Terra through the Codex CLI where it is available. The phase text is the core of the sub-agent's brief; add current tree state, relevant conventions, return format and the files the sub-agent owns.

Before execution, fill missing context or task-critical detail in the plan, with concrete steps, paths and commands where the work's uncertainty or risk warrants them. Let the executor investigate evidence that contradicts a plan assumption. Preserve the goal and decisions; seek input only for a material change beyond the user's authorization. Tell the user about every substantial or consequential change you make to the approved plan, whether it fills a gap or follows from an investigation; minor ones need no mention. A model change alone does not require rewriting a sufficient approved phase.

  • Update the checklist to [~] before dispatching
  • The sub-agent (or you, from its result) marks [x] or [!] on completion

5. Verification

After completing all checklist items, run /doublecheck before marking the task complete. It briefs fresh-context reviewers rather than relying on your own recollection, and it consistently catches problems the executor missed.

6. Task File Disposition

Apply only to TASK_*.md files created by this skill, never to existing PLAN files.

After /doublecheck passes:

  • Delete the TASK file if it contains only transient execution tracking, routine verification, and a completion summary.
  • If it records durable findings, decisions, constraints, or follow-ups, propose an appropriate destination based on repository conventions.
  • After approval, preserve the information there and delete the TASK file, unless archiving the file itself is appropriate.
  • Keep the file while work is incomplete or blocked.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

Reviews can't be read right now.

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.