--- name: "monthly-post-plan" description: "Produce a fortnight of social posts: read the content log first, batch-write all posts of one type together from a controlled source file, build each visual from the last approved design, then record what changed so the corrections stick." --- # Monthly Post Plan Produces one fortnight of posts for a person publishing their own content — not a team running a brand account. The difference matters: a reader notices when one voice slips, and does not notice when a brand's does. Three stages, about an hour in total. **Read the log first, write it last.** A run that produces posts and does not update the log has not finished — the next one will repeat content and remake corrections already made. ## Set these up once Two files live outside this skill and are read by every run. Create them on the first run by interviewing the user, then never ask again. | File | What it holds | | --- | --- | | **Setup document** | Who is writing, the voice rules, the two reference pieces, the pillars and quotas, the rhythm, the structure of each post type, and the one rule about what may never be invented. | | **Content log** | What was published and scheduled, which design is each type's current template, the numbered style rules, rejected hooks, spent material and open gaps. | Placeholders to fill on the first run: | Placeholder | What it is | | --- | --- | | `{{YOUR_NAME}}` / `{{BRAND_NAME}}` | as they appear in captions and file names | | `{{SOURCE_FILE}}` | the one file holding real anecdotes, cases and figures | | `{{LOG_LOCATION}}` | where the content log lives | | `{{DESIGN_TOOL}}` / `{{DESIGN_ACCOUNT}}` | where visuals get built | | `{{SCHEDULER}}` / `{{TIMEZONE}}` | where posts get scheduled | | `{{PILLARS}}` / `{{QUOTAS}}` | subjects and how many of each per month | | `{{RHYTHM}}` | which day carries which post type | | `{{LANGUAGE}}` | output language, and any post type that runs in a second one | Never put credentials in either file. Say where they live. ## Stage 1 — Take stock (10 min, no writing yet) 1. Load the voice and visual rules **first**. A run that starts by writing produces copy the user then rewrites by hand. 2. Read the content log in full. 3. Mark each day in the fortnight: published · scheduled · visual ready · copy written · empty. Fill only the empty ones. 4. Flag anything that contradicts the plan — a post breaking the pillar quotas, a theme repeating — rather than quietly correcting it. Only the user knows what they posted or scheduled by hand. Ask, rather than assume the log is complete. ## Stage 2 — Write the fortnight in batches (25 min) 1. Write all posts of one type together: series episodes in their set order, statement cards in one batch, the weekly pillar in another. 2. Before each piece, read the **two** finished reference pieces in the setup document. One example is not a pattern — with one, you copy the surface of it. 3. Take every anecdote and every figure from `{{SOURCE_FILE}}`. Nothing from memory. Nothing merely plausible. 4. Where that file does not cover a post, **leave the slot empty and marked, and write the question that would fill it.** An invented example reads better than an empty slide, which is exactly why it gets through. 5. Every post leaves this stage with its final copy written — not with a brief. ## Stage 3 — Build the visuals (25 min) 1. For each type, open its **current template: the last post of that type the user approved.** Not the first one designed. If they corrected a post after it was built, the corrected version is the template. 2. Copy it and replace the text. Never generate a design from scratch — a new design starts from zero corrections, however good the prompt. 3. Rewrite an existing text element rather than adding a new one; a new box arrives without the brand font and nobody spots it until it is published. 4. Keep the sentence and drop the font size. Shortening the line that was carrying the meaning costs more than a smaller font ever does. 5. Trust the text layer, not the thumbnail — the preview can lag an edit that already worked. 6. Update the template table in the same run. The post just made becomes its type's template. 7. Deliver the design links **grouped by post type**, not by date. The user opens all the statement cards at once, not one per day. ## After the run — record what changed This is the stage that makes the second run better than the first. For each correction the user made, decide which it is: - **A one-off fix** — a word changed in one particular context. It goes nowhere. - **A pattern** — a choice they would make again, in any text. This gets written into the log. The test: *if you can phrase the change as an instruction for a future text, without mentioning the post it came from, it is a pattern.* Write each rule as an instruction with its example beside it — "Unpack nouns into verbs: *the effort you feel*, not *felt effort*." Number the rules consecutively and never renumber them. **Two occurrences make a rule.** One gets recorded as a single occurrence and waits for the next post to confirm it. Don't inflate a rule out of one case — an over-general rule is its own kind of drift. Then write back: what was published, which hooks were used, the open gaps and their questions, the designs awaiting review, and any figure that went in without a source. ## Don'ts - Don't fill a gap with an invented example, a plausible client, or a figure you cannot source. - Don't build a visual from the original template, or from scratch. - Don't batch a whole month. A month is a long time to be wrong; a rolling fortnight stays correctable. - Don't finish a run without updating the log. --- ## Appendix A — Setup document template ``` # Monthly Post Plan — setup ## 1. Who is writing Name and brand name, as they appear in captions and file names. The role written from, in one line. ## 2. Voice, in rules not adjectives "Warm but professional" is not a rule. A rule is something a draft can fail. - Three expressions or constructions refused, and why - Punctuation not used - One sentence that is exactly right, and one that is close but wrong - Keep these numbered. Every new rule gets the next number and the example that produced it. ## 3. Two complete reference pieces Two finished posts, pasted in full. Not a description of the style. ## 4. Pillars and quotas Subjects, and how many of each per month. Which are mandatory, which are flexible, and any cap — how often a lead magnet may be promoted, how often a second language may be used. ## 5. Rhythm Which day carries which post type. "Three times a week" is an average, not a rhythm. ## 6. Post types, and the structure of each For each: how many slides or lines, what goes first, what goes last, how it closes. ## 7. What may never be invented The most important section. Name the file holding the real material. Then write the rule: nothing comes from outside that file; uncovered slots stay empty and marked, with the question that would fill them. ## 8. Tools and where things live Where finished copy is saved, where the log lives, the design tool and account, the scheduling tool, account and time zone, the output language. No credentials — say where they live. ## 9. The template rule For each post type, the current template. The answer is always the last post of that type that was approved. Write the table anyway and update it every run. ## 10. After the run: what was corrected One-off fix, or pattern? Two occurrences make a rule. ## 11. After the run: what was left The marked gaps and their questions, the designs awaiting review, what was not covered and why. ``` ## Appendix B — Content log template ``` # Content log ## 1. Published | Date | Type | Subject | Hook used | Link | The hook column matters as much as the subject: a spent hook coming back is harder to spot than a spent topic. ## 2. Scheduled | Date | Type | Subject | State | Link | State is one of: copy written · visual ready · scheduled. ## 3. Current templates | Post type | Current template | Produced | Approved | The template for a type is the last post of it that was approved. If a post was corrected after it was built, the corrected version goes here. ## 4. Style rules Numbered, consecutive, never renumbered. Each rule gets the example that produced it. A rule with one occurrence is marked as such and waits. ## 5. Writing benchmark Construction patterns, not rules. What goes on the second slide, what a close looks like, what a swap label has to say. ## 6. Rejected hooks | Hook | Rejected because | ## 7. Spent material Anecdotes and stories already used, with where. A good story used twice in four months reads as a thin practice. ## 8. Open gaps | Date | Type | What is missing | The question that would fill it | ## 9. Designs awaiting review Designs that inherited a background photograph from the post they were copied from. ## 10. Unestablished sources Figures that went into a post without a source being established. ``` --- Built by Laura Balan, 3rd Space AI Circle. Card: https://3rdspacebucharest.com/ai-workflows/monthly-posts/