agentleFS
CAFEERON21br18CAFEERON21br18/Daily_Master_Prompt/SKILL.md

create

Skill1 starsChanged 23 days agoMITView on GitHub

Use when the user wants a daily master prompt, a recurring or unattended work-session prompt, or a "pick up where we left off every day" routine for a project, or wants to change or add an addendum to an existing DAILY_PROMPT.md.

19 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use compare, publish and 3 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.
  • n24q02mSkill

    compare

    Structured comparison of 2+ alternatives with consistent criteria and decision matrix

    Also in 3 projects that use this

  • VercelSkill

    publish

    Publish a single demo to v0 and optionally create Typefully drafts. Use when the user says "publish", "publish to v0", or "create typefully draft" for a demo.

    Also in 4 projects that use this

  • genomewalkerSkill

    tasks

    Browse and resume tasks, threads, and background jobs across sessions

    Also in 3 projects that use this

  • AnthropicSkill

    analyze

    Answer data questions -- from quick lookups to full analyses. Use when looking up a single metric, investigating what's driving a trend or drop, comparing segments over time, or preparing a formal data report for stakeholders.

    Also in 3 projects that use this

  • AnthropicSkill

    status

    Case status summary by audience — client-facing (plain language), internal (for the professor), or court-ready (formal caption format per local rules). Same facts, different framing and depth. Use when a student needs to update the client, brief the professor, or prepare a court status report.

    Also in 3 projects that use this

The file

We split this skill into the rules it gives an agent

4 rules. 0 are common, 1 is distinctive, and 3 are very specific to this project.

  • 0

    Common

    Most of the 17 edited versions we read keep these.

  • 1

    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:

    All in the project folder (<folder>), built from the templates in ${CLAUDE_SKILL_DIR}/templates/:

    line 21
  2. Distinctive:

    Find the folder: Use the folder in the arguments.

    line 36
  3. Very specific:

    Check: grep -n "{{" <folder>/*.md returns nothing.

    line 41
  4. Very specific:

    When DAILY_PROMPT.md already exists, read it and every state file first.

    line 52

“Common” is measured: kept by at least 6 in 10 of the 17 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

Create a Daily Master Prompt

Overview

A daily master prompt is one self-contained file, DAILY_PROMPT.md, that any fresh session can execute to pick up a project, do a day of work, and leave it ready for the next day. The project's memory lives in five state files next to it, never in the conversation. After the first run the user only types:

Read <folder>/DAILY_PROMPT.md and execute it.

Input from the user: $ARGUMENTS (a folder and/or an explanation of the project; either may be missing).

What exists when you finish

All in the project folder (<folder>), built from the templates in ${CLAUDE_SKILL_DIR}/templates/:

FileHoldsChanges
DAILY_PROMPT.mdMission, milestone, rules, context map, session procedureOnly when the user asks
BACKLOG.mdOrdered work items with IDs and statusEvery run
DEVLOG.mdAppend-only dated log of each runEvery run
HANDOFF.mdCurrent state, next 3 actions, warningsRewritten every run
QUESTIONS.mdQuestions for the user, with a recommendation eachRuns ask, user answers
APPROVALS.mdRisky or irreversible actions waiting for yes/noRuns request, user decides

Progress goes in the state files. DAILY_PROMPT.md holds only what stays true for weeks.

Steps

  1. Find the folder. Use the folder in the arguments. Otherwise use the folder holding the project's spec or code; if several fit, ask.
  2. Gather context by reading, not asking. The user's explanation; README*, CLAUDE.md, AGENTS.md; specs (GDD, PRD, roadmap, docs/); git log --oneline -30 and git status; a 2-3 level file tree; manifests and scripts that reveal build, test, lint and run commands; connected MCP servers the work needs; TODO/FIXME notes; any existing state files. Separate what is done from what is not.
  3. Ask only for what the files don't say, in one AskUserQuestion call (max 4 questions), from this list: current milestone and its definition of done; how a finished item is verified; hard rules (never publish/deploy, ask before new dependencies, branch and push policy); whether runs are attended, unattended, or both. If no one answers, take the safest reading (unattended, no push, no publish) and record each assumption in QUESTIONS.md with its "Assumed until answered" line.
  4. Write DAILY_PROMPT.md from DAILY_PROMPT.template.md. Every section names this project's real paths, commands, tools and milestone. In the daily line, write the folder relative to the repository root (where the user opens Claude), not an absolute path; when the folder is the root itself, the line is Read DAILY_PROMPT.md and execute it. If running or testing the project could touch real data (user files, live accounts, production, a published place), add a ground rule and a verification method that use a temporary or test copy instead.
  5. Create the state files from their templates. If one already exists, merge into it; never overwrite or reorder the user's history. Seed BACKLOG.md from the milestone, spec gaps, TODOs and the user's goals, ordered so item 1 can start today and each item fits in one session, each with a concrete "Done when". Seed HANDOFF.md with the starting state. A template whose example entry doesn't apply yet keeps its header and format lines, with None yet. in place of the entry. Check: grep -n "{{" <folder>/*.md returns nothing.
  6. Commit the new files if the folder is in a git repo (chore: add daily master prompt), following the repo's commit rules.
  7. First run. Unless the user said not to, execute DAILY_PROMPT.md now, exactly as a fresh session would. This is the prompt's test: when a step was ambiguous, failed, or needed knowledge that isn't in the files, fix the prompt and note the fix in DEVLOG.md.
  8. Hand off. Tell the user what was created, what the first run did, and give the daily line in a code block:
    Read <folder>/DAILY_PROMPT.md and execute it.
    Mention /daily-master-prompt:run <folder> as the same thing, and that the line also works as the prompt of a scheduled task.

Updating an existing prompt

When DAILY_PROMPT.md already exists, read it and every state file first. Apply the change in place, or under ## Addenda as a dated entry when the user asks for an addendum. Keep all state. Log the change in DEVLOG.md.

Common mistakes

MistakeResultDo instead
Generic prompt ("work on the project")Each run rediscovers the projectReal paths, commands, milestone, done-definition
Progress written into DAILY_PROMPT.mdPrompt goes stale and contradicts state filesProgress lives in BACKLOG/HANDOFF/DEVLOG
No verification step"Done" items that don't workA concrete check per item (tests, build, playtest, lint)
Runs stop at the first blockerUnattended days produce nothingLog in QUESTIONS/APPROVALS, continue with the next item
Backlog items too bigHalf-finished work every daySplit until one item fits one session
Overwriting existing state filesUser loses historyMerge

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.