Linkedin Content Planner

LinkedIn Content Planner

Posting without a plan produces a recognisable pattern: three posts in an enthusiastic first week, all on the same subject, then silence, then a product announcement to an audience that has stopped expecting anything. The individual posts may be good. The sequence is not, because nobody decided in advance how often each subject should come up, which posts need production time, or where the material for week three would come from.

This skill makes those decisions once, on paper. The agent turns a small config (pillars with target shares, posting days, a format mix, a few limits) into a dated plan in which pillars and formats are spread evenly, each slot states the one thing it asks of the reader, and every slot is either tied to a real piece of material or visibly marked as lacking one. Imbalance is flagged before the first post is written, when it is still cheap to fix.

Offline only. The skill produces a plan for a person to act on. It does not post, schedule, read a feed, pull analytics or call any API. Inputs are a config file and, optionally, a story bank file, both supplied by the user.

Scope boundary. This skill decides what goes on which date; it does not write the posts. Drafting a slot is linkedin-post-writer; cleaning a draft's phrasing is linkedin-humanizer; testing its opening line is linkedin-hook-analyzer. Finding the material a slot needs is linkedin-story-interviewer, which also owns the story bank format this skill can optionally read. Adapting an existing talk or article into a slot is linkedin-content-repurposer. The plan reserves conversation days but what to say on them is linkedin-comment-writer, linkedin-reply-manager and linkedin-thread-tracker. Profile copy is linkedin-profile-optimizer, multi-person programmes are linkedin-employee-advocacy, and deciding from past results what to change next month is linkedin-engagement-analytics.

When to use this skill

• "Plan my week" or "what should I post this month"
• A person posts in bursts and wants a cadence they can keep for a quarter
• A launch, fundraise or hiring push is coming and needs to sit inside normal posting rather than replace it
• A founder is writing for investors, candidates and buyers at once and the feed has become all product
• A story bank exists and nobody has checked whether it can feed the intended cadence
• An existing calendar feels repetitive and needs an objective look at its mix

Inputs the skill expects

• A start date and a length of one to thirteen weeks
• Posting weekdays, and the weekdays reserved for conversation on other people's posts
• Three to five pillars, each with a target share, the reader ask it usually carries, and the formats it suits
• A format mix with shares, marking which formats need production time
• Limits where the defaults do not fit: posts per week, pillar cap, promotional posts per week, produced posts per week
• Optional: dates to skip, and a story bank file whose pillar names match the plan's

Clarify First

Before building the plan, confirm these inputs. If any is unknown or vague, ASK — do not assume:

• General or founder pillar set — a founder writing for investors, hires and buyers needs pillars keyed to those readers; the general set is keyed to subject matter. Choosing wrong produces a plan that is balanced on paper and aimed at nobody.
• How many posts a week the person will still be making in week ten — the plan is built on this number; an optimistic answer creates slots that go unfilled and a pattern of gaps.
• How much production time exists per week — carousels, video and designed images take hours that text does not; this caps the format mix regardless of preference.
• Whether a story bank exists — with one, slots are bound to entries and shortages show up as flags; without one, every slot is a topic with nothing behind it yet.

Stop rule: ask only the two or three that most change the plan. If the user says "just plan it", use the general set at three posts a week, text-heavy, and list those assumptions at the top of the plan.

Workflows

Quick start: copy assets/plan_config_template.json, set the dates and pillars, and run Workflow 1.

Workflow 1 — Build a month plan from a config

1. Agree pillars and shares using references/pillars-and-mix.md. Shares must sum to 100; the builder refuses to guess otherwise.
2. Set posting and conversation weekdays, skip dates, and the format mix. Mark formats that need production with "produced": true.
3. Run the builder. Read the flags before the calendar.
4. Resolve every error flag by changing the config, not by hand-editing the output. Decide deliberately about each warn.
5. Copy the result into assets/weekly_plan_template.md and add a one-line angle per slot.

python3 tools/linkedin/linkedin-content-planner/scripts/plan_builder.py \
  --input tools/linkedin/linkedin-content-planner/assets/sample_plan_config.json

Workflow 2 — Bind the plan to a story bank and find the shortfall

1. Check that the bank's pillar names match the plan's exactly.
2. Run the builder with --story-bank. Each slot takes the next ready entry filed under its pillar, unused entries first.
3. Read every CP-012 flag: those are slots with no material behind them.
4. For each short pillar, either lower its share or schedule an interview session before the plan starts. Do not leave the slot to be improvised.
5. Validate the final version with --fail-on error so a broken mix cannot be handed on.

python3 tools/linkedin/linkedin-content-planner/scripts/plan_builder.py \
  --input tools/linkedin/linkedin-content-planner/assets/sample_plan_config.json \
  --story-bank tools/linkedin/linkedin-content-planner/assets/sample_story_bank.json \
  --fail-on error

On the sample this exits 0 with four warnings: the bank cannot cover the offer pillar at all, and the format mix asks for more produced posts than one a week allows.

Workflow 3 — Plan a founder fortnight

1. Read references/founder-pillar-set.md and confirm the three readers (investors, candidates, buyers) are the ones the founder actually wants.
2. Start from assets/sample_founder_config.json or set "pillar_set": "founder" in a config with no pillars key.
3. Build a two-week plan first. Founders' weeks change too fast for a month to survive contact.
4. For each slot pick an angle from the founder reference and name the real event behind it. A slot with no event is cut, not padded.

python3 tools/linkedin/linkedin-content-planner/scripts/plan_builder.py \
  --input tools/linkedin/linkedin-content-planner/assets/sample_founder_config.json \
  --format json

Exit-code contract. 0 plan built; 1 a flag reached the --fail-on level; 2 the config or bank is missing, not JSON, or invalid (shares not summing to 100, unknown weekday, bad date, no slots).

Decision frameworks

Which pillar set

Cadence

All figures here are planning heuristics, not platform rules. Platform behaviour changes; treat anything about reach as current as of writing and verify it against your own results.