agentleFS
kabo-shkabo-sh/kabo-mcp-plugin/skills/plan-content-calendar/SKILL.md

plan-content-calendar

Skill2 starsChanged 20 days agoMITView on GitHub

Build a reviewable four-week or date-range content calendar (up to 91 days), or propose which already-selected item fits one exact publish slot, from selected ideas, script packages, production state and publish-timing results. Also CSV/ICS export and reminder or publish-preparation handoffs. Use for planning when to post what, or questions like which one to post at 4pm today. Not for research, choosing new ideas, rewriting content, inventing a best time, uploading or publishing.

4 other projects use a version of this.

Adopted by

Most recently taken up by

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 4 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. Distinctive:

    When the host supports one, run an artifact-backed calendar in a foreground sub-agent and wait for its final result before replying.

    line 26
  2. Very specific:

    Never skip the request validator, the deterministic replay check, the final validator or the projection hash checks.

    line 111
  3. Very specific:

    Every scheduled slot shows its complete title, audience problem, promise, angle and opening Hook, never shortened to #2.

    line 128
  4. Very specific:

    Do not use web search, another Skill, model memory, generic platform advice, a different artifact, or generated replacement content to make …

    line 140

“Common” is measured: kept by at least 6 in 10 of the 4 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

Plan Content Calendar

Turn immutable upstream decisions into a deterministic, reviewable execution plan. content-calendar.v1 JSON is the product truth; Markdown, CSV and ICS are projections of it.

Delivery. For an artifact-backed run, the renderer is the final step and its deliverable is content-calendar.md. Read it whole and reply with it as the calendar body, once, in the creator's language, ending with creator_report: content-calendar.md; never regenerate a second schedule in prose. Keep the planned and target counts, every unfilled slot and conflict, production conditions and action boundaries. With it, give the validated content-calendar.json path and its status, the CSV and ICS paths when exported, unresolved decisions and gaps, and a plain-language summary of each proposed calendar, reminder and publish-preparation action with its exact slot scope. State plainly that every proposed action needs the creator's separate confirmation and that no calendar, reminder, upload, schedule or publication was created because an artifact or ICS file exists. Never print snake_case action enums or internal field paths unless the user asks for the technical contract.

When the host supports one, run an artifact-backed calendar in a foreground sub-agent and wait for its final result before replying; never run it in the background and tell the user the report will come later. Without a foreground sub-agent, follow the same steps in the main thread.

Route the request first

The request may arrive in any language. Decide which of these it is before anything else:

The requestWhat to do
Strategy or workflow advice (sustainable filming/editing, restart guidance, content-role mix, a learning roadmap), a review of times or reminders the creator already stated, or a proposed plan for topics the creator suppliedAnswer in the conversation, per references/advice-and-proposals.md. Reserve no run and call no script: without selected-idea artifacts the pipeline can only fail.
A four-week plan, an explicit date range, CSV/ICS export, or which selected item fits one exact slotThe artifact-backed pipeline below. Pick the mode in references/request-modes.md.
New ideas, holiday research or content selection a calendar would needThat is another Skill's job; materialize the slots only after selection. Missing selections stay explicit unfilled slots.

For those conversation routes, read references/advice-and-proposals.md first, then answer now:

  • Topics the creator supplied, to be given angles, formats and posting dates: that request authorizes a proposal. Deliver the complete proposed calendar in this reply, every supplied topic kept, with a sensible default cadence labelled as an assumption. Title it an unvalidated draft, not a native calendar, and name what would make it one (the selected items and the account). An unknown platform, start date or timezone does not hold it back: give the plan in relative order (week 1 day 1, …) and ask only for the missing anchor; never invent an exact instant, and never reply with questions alone.
  • Advice: answer the strategy now with stated assumptions, not a routing menu; missing account data blocks only an account-specific diagnosis.

A four-week horizon does not imply eight posts: keep the creator-selected content set, never expand or clone it to fill weeks, and report unused slots honestly.

Clarify once, then fail closed. For an artifact-backed calendar, ask one consolidated question when a required target, account, date, creator timezone, selected item, cadence or source artifact is missing, after first delivering any advice or review the information already supports. Relative times such as "today" or "4pm tomorrow" become an offset datetime in the creator's IANA timezone; never read the machine timezone or clock, and ask when an exact instant needs an unknown timezone.

Running commands

<skill directory> is the folder holding this SKILL.md. Reserve exactly one run per artifact-backed request:

python3 '<skill directory>/scripts/static_host.py' init --workspace '<user-workspace>'

It prints the run directory, <run> below. Copy every upstream artifact the calendar binds (SelectedIdeas, ScriptPackage requests and packages, publish-timing request, history, recommendation and schedule, final media and covers) into <run> unchanged with cp, because every ref is a path relative to the run. Every command starts with cd '<run>' && . Write authored JSON with a single-quoted heredoc or the host's file tool. Every command here is complete: do not probe scripts with --help or read their source, and never edit the scripts, schemas or validators. Quote paths and use grep rather than rg. A failed step stops the run: report the error it printed; never fabricate artifacts or confirmations to pass it.

Pipeline

  1. Build content-calendar-request.json per references/request-building.md (refs, items, joins, timezone provenance, production progress and publication assets), then fill every ref's id and hash from the files:

    cd '<run>' && python3 '<skill directory>/scripts/calendar_ref.py' content-calendar-request.json
  2. Validate the request, then build the raw calendar:

    cd '<run>' && python3 '<skill directory>/scripts/validate_content_calendar.py' request content-calendar-request.json --run-root .
    cd '<run>' && python3 '<skill directory>/scripts/build_content_calendar.py' content-calendar-request.json --run-root . --output content-calendar.raw.json
  3. Export CSV and ICS from that same JSON (this only writes local files and their hashes; it creates no reminder and saves no calendar). For a reviewable draft without export, copy content-calendar.raw.json to content-calendar.json instead and continue:

    cd '<run>' && python3 '<skill directory>/scripts/export_content_calendar.py' content-calendar.raw.json content-calendar-request.json --run-root . --csv content-calendar.csv --ics content-calendar.ics --output content-calendar.json
  4. Validate the final calendar and its projections, then render:

    cd '<run>' && python3 '<skill directory>/scripts/validate_content_calendar.py' calendar content-calendar.json content-calendar-request.json --run-root .
    cd '<run>' && python3 '<skill directory>/scripts/render_content_calendar.py' content-calendar.json content-calendar-request.json --run-root . --output content-calendar.md

Never skip the request validator, the deterministic replay check, the final validator or the projection hash checks. The planning rules the builder applies, and that the report must keep, are in references/planning-rules.md.

Ownership boundaries

  • Ideas own what to make: only selected-idea.v2 can occupy a slot; idea-set.v3-candidate1 is backlog and stays candidate_requires_selection.
  • Script & Packaging owns how to make it: copy approved package fields; never draft missing headlines, captions, scripts, covers or assets here.
  • Publish Timing owns evidence about when: preserve its lineage; never recompute a recommendation or turn no_clear_window / abstain_* into a generic best time.
  • Runtime owns confirmation and durable actions: calendar writes and reminders happen only after the creator confirms each scoped action; platform media preflight, upload, scheduling and publishing belong to an authorized connector, never to this Skill.
  • Trial Reels, variants, frequency experiments and post-performance analysis are LATER: add no experiment fields and never present a normal slot as an experiment.
  • Every scheduled slot shows its complete title, audience problem, promise, angle and opening Hook, never shortened to #2; observed production state and structural publish preparation are shown separately, and neither stands in for the other.
  • The highest state this Skill owns is ready_for_review. An upstream artifact may truthfully say a slot was already accepted or scheduled; preserve that state, never claim this Skill caused it.

Hard product contracts, kept verbatim:

  • never choose on the creator's behalf
  • No generic platform best-time fallback
  • No upload, schedule, publish or external write side effect

Do not use web search, another Skill, model memory, generic platform advice, a different artifact, or generated replacement content to make an invalid request pass. The full fail-closed list is at the end of references/planning-rules.md.

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.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

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.