write-draft
Skill0 starsChanged 58 days agoMITView on GitHub
Use when the user asks to draft a LinkedIn post from a selected idea or a Story Bank Post Spine, run the LinkedIn Writer/Post Writer, or explicitly invokes /write-draft. Converts a status:selected Idea Note (default path) OR a Content-Learnings/story-bank.md Post Spine (--spine <id>) into a post draft in Drafts/, picking a hook formula from Content-Learnings/hook-formulas.md and optionally a founders angle from Content-Learnings/founders-angle-library.md by engagement goal, grounded in linked research (idea path) or the cited Story Bank entries (spine path), following Content-Learnings/voice-guide.md. Does not assign posting days (/plan-week) or handle approval/scheduling.
1 other project uses a version of this.
Adopted by
Most recently taken up by
- calven-ai/marketing-as-codetheir own version, first seen Oct 7, 2026
Other versions1 project uses a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 1
- calven-ai/marketing-as-code★ 10same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository38 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- audit-draftSkill
- check-plagiarismSkill
- critique-draftSkill
- critique-draft-substack-articleSkill
- critique-draft-substack-noteSkill
- critique-draft-xSkill
- draft-commentSkill
- draft-replySkill
- extract-hookSkill
- generate-ideasSkill
- generate-visualSkill
- generate-visual-substack-articleSkill
- generate-weekSkill
- generate-week-substackSkill
- generate-week-xSkill
- humanize-draftSkill
- interviewerSkill
- monitor-engagementSkill
- optimize-profileSkill
- plan-advocacySkill
- plan-weekSkill
- plan-week-substack-articleSkill
- plan-week-substack-noteSkill
- plan-week-xSkill
- publish-substackSkill
- pull-analyticsSkill
- pull-analytics-xSkill
- repurpose-postSkill
- research-topicSkill
- review-draftsSkill
- schedule-approvedSkill
- schedule-approved-xSkill
- update-playbookSkill
- write-draft-substack-articleSkill
- write-draft-substack-noteSkill
- write-draft-xSkill
The file
Show its rules and text
We split this skill into the rules it gives an agent
17 rules. About 0 are common, 6 are distinctive, and 11 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈6
Distinctive · estimate
This author's own choices. Worth a look.
11
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 53
Idea path: read the Idea Note and every research note in its sources[].
- Very specific: line 57
Spine path: read the spine row plus every Story Bank row it cross-references (by receipt_id/position_id).
- Distinctive: line 68
Never invent a sixth value.
- Very specific: line 71
Read Content-Learnings/voice-guide.md in full before drafting.
- Very specific: line 83
Read Content-Learnings/hook-formulas.md fresh each run — never rely on a cached memory of it, since /extract-hook and human promotions grow …
- Very specific: line 100
Read Content-Learnings/founders-angle-library.md fresh each run.
- Very specific: line 101
Look for a canonical angle (never proposed) whose bracketed template slots can genuinely be filled from an actual …
- Distinctive: line 122
If a founders angle was used (step 6), it shapes the whole structure of the post per its template — the angle's bracketed narrative arc …
- Distinctive: line 127
If no founders angle is in play: hook → the substantive content → an optional closing discussion question, only if the voice guide's CTA …
- Very specific: line 130
If the idea's format is opinion (or the content otherwise includes a subjective take, not just what the research/Story Bank supports), …
- Distinctive: line 140
Follow the voice guide's "avoid" list explicitly — check the draft against it before finalizing, don't just aim for the tone and hope.
- Very specific: line 168
Leave viral_score at 0 and visual_ids empty — scoring the draft and generating visuals are later phases (Quality/Critic Agent, Visual …
- Distinctive: line 194
Never invent a fact, statistic, or quote not present in the linked research notes (idea path) or the cited Story Bank rows (spine path).
- Very specific: line 198
Never draft an idea that isn't status: selected (run /plan-week first).
- Very specific: line 200
Never invent a founder anecdote or number that isn't present in Content-Learnings/story-bank.md — no genuine match means no founders angle …
- Very specific: line 203
Never select a proposed-status hook formula or founders angle for drafting — both must be canonical.
- Distinctive: line 211
Never mark anything as approved, scheduled, or published — this skill only produces a draft file.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Write Draft / Post Writer (LinkedIn Writer — Phase 4b, reworked Phase 17)
Converts either a status: selected Idea Note (default, primary path) or a
Content-Learnings/story-bank.md Post Spine (--spine <id>, new path) into
an actual LinkedIn post draft, written to Drafts/ using
_Templates/Draft-Note.md. Also selects a hook formula from
Content-Learnings/hook-formulas.md by engagement goal, and — for Career /
Interview Prep categories, or whenever asked — an optional founders angle
from Content-Learnings/founders-angle-library.md, grounded in real Story
Bank material. Assumes /plan-week has already assigned a day to the idea
(idea path only — a spine has no day assignment to assume).
Arguments
- idea-id (optional, default path) — draft one specific idea. If
omitted, and no
--spineis given, draft every idea currentlystatus: selectedthat doesn't already have a draft linked to it. - --spine <spine_id> (optional, new path) — draft from a Post Spine
row in
Content-Learnings/story-bank.md's## Post Spinestable instead of an Idea Note. Mutually exclusive with idea-id. - --goal <goal> (optional, either path) — explicit engagement goal
override. Closed list:
likes,comments,shares,saves,profile-visits. - --founders (optional, either path) — force-consider a founders angle even outside Career / Interview Prep categories. Still subject to the same real-Story-Bank-match-or-skip rule below — this flag only widens when the attempt happens, never lowers the bar for using one.
Process
1. Determine the entry path
- idea-id given, or no
--spine(default): the existing idea-based flow below, unchanged in spirit — this stays the primary path. --spine <spine_id>given: load that row fromContent-Learnings/story-bank.md's## Post Spinestable byid. Itsreceipt_id/position_idcross-references, if set, are the grounding for this draft — treat them with the exact same never-fabricate discipline as an idea's linked research notes: every factual claim in the draft must trace back to the spine's own fields (hook_angle,story_beat,position,suggested close) or the Story Bank row(s) it cross-references. If the spine sets noreceipt_id/position_id, the draft must not invent one — work only from what the spine itself states. There is no idea for this path, so there's no idea status to flip at the end; instead the Draft Note recordsspine_id(see step 7) so its provenance is traceable.
2. Load context (idea path) / load the spine (spine path)
- Idea path: read the Idea Note and every research note in its
sources[]. All factual claims in the draft must trace back to something in those research notes' Summary/Key Findings — never introduce a new statistic, quote, or claim that isn't already there. - Spine path: read the spine row plus every Story Bank row it
cross-references (by
receipt_id/position_id). Those rows are this draft's entire factual grounding — there is no research note to fall back on.
3. Determine the engagement goal (both paths)
Precedence: explicit --goal <goal> argument > the idea's own goal field
if it has one and is set (idea path only) > a Content-Learnings/
playbook.md-evidenced pattern for this category (real post ids attached
as evidence), if one exists > default comments. Closed list, exactly
these five values: likes, comments, shares, saves, profile-visits.
Never invent a sixth value.
4. Read the voice guide and playbook
Read Content-Learnings/voice-guide.md in full before drafting. This is
the current source of truth for hook style, structure, length, tone, and
things to avoid — it starts as a documented default and will be updated by
later phases as real approval/edit data accumulates. Always re-read it
fresh each run; don't rely on a cached memory of it, since it can change.
Also check Content-Learnings/playbook.md for any evidenced rule about
length, hook style, or hashtags for this category/content_type. If one
exists (with real post ids attached as evidence), prefer it over the
voice guide's generic defaults for that specific dimension.
5. Select a hook formula
Read Content-Learnings/hook-formulas.md fresh each run — never rely on a
cached memory of it, since /extract-hook and human promotions grow it
over time. Filter to status: canonical rows only (never select a
proposed row — those aren't trusted for drafting yet) tagged for the
chosen engagement goal; a formula with no engagement_goals value is
general-purpose and eligible for any goal. Among the eligible set, prefer
a formula not used in the last ~5 drafts — check recent Drafts/ notes'
hook_formula field (new frontmatter field, see step 7) to avoid
repetition. If every eligible formula was used recently, repetition is
acceptable rather than forcing a worse goal-fit; goal match beats novelty.
6. Optional founders angle
Triggered automatically if the idea's (or spine's) category is Career
or Interview Prep, or if the user passed --founders. Otherwise
skipped without comment (it's simply not applicable this run).
If triggered:
- Read
Content-Learnings/founders-angle-library.mdfresh each run. - Look for a
canonicalangle (neverproposed) whose bracketed template slots can genuinely be filled from an actualContent-Learnings/story-bank.mdentry (a Receipt, Turning Point, Scar, or Defensible Position) matching the topic. - Real fit found → use it. Fill every bracket with the real Story
Bank value, nothing invented or rounded, and cite that Story Bank row's
idin the Draft Note'ssources/founders_angle note. - No real fit → drop the founders angle entirely for this draft, silently in the draft text (don't leave a half-filled template) but not silently in the report — call this out explicitly in "Report back" below. Never invent a plausible-sounding number or anecdote just to complete a bracket.
7. Draft the post
- Hook: idea path uses the idea's
hookfield, refined if needed and shaped by the selected hook formula's opening mechanic; spine path uses the spine'shook_angle, similarly shaped. The hook formula shapes the opening line only — it never overrides step 9's opinion/fact separation, and it must never misrepresent certainty (e.g. don't let a "contrarian claim" formula state a personal opinion with the same declarative confidence as a sourced fact). - If a founders angle was used (step 6), it shapes the whole structure of the post per its template — the angle's bracketed narrative arc becomes the post's body, still filled only with real Story Bank specifics and still following the same fact-discipline as everything else here.
- If no founders angle is in play: hook → the substantive content → an optional closing discussion question, only if the voice guide's CTA guidance supports one here.
- If the idea's
formatisopinion(or the content otherwise includes a subjective take, not just what the research/Story Bank supports), frame it explicitly as a view — "I think," "the mistake I keep seeing," "my take" — rather than stating it with the same declarative confidence as a sourced fact (REQUIREMENTS.md §21: clearly distinguish opinions from facts). Sourced claims and personal opinions should read differently to the reader, not blend together. - Length: per the voice guide's default (~1,200–1,500 characters, comfortably under LinkedIn's ~3,000 limit) unless the material genuinely needs more room.
- Follow the voice guide's "avoid" list explicitly — check the draft against it before finalizing, don't just aim for the tone and hope.
8. Hashtags
3–5, directly tied to topic/category, no stuffing. Unchanged from before.
9. Write the Draft Note
Copy _Templates/Draft-Note.md into Drafts/ as
YYYY-MM-DD--kebab-slug.md (idea path: use the idea's target_date, not
today, if the two differ; spine path: use today's date, since a spine has
no target_date — the filename should reflect when it's meant to post).
Fill in: category, format, hook_style, hashtags, status: draft,
history: [{action: created, date: today, note: ...}], plus the new
frontmatter fields:
hook_formula: the matchedformula_idfrom step 5 (e.g.F10).engagement_goal: the goal determined in step 3.founders_angle: the matchedangle_idfrom step 6 if one was used (e.g.A5), else left empty.- Idea path:
idea_idset to the idea's id,spine_idleft empty.sourcesinherited from the idea (research note ids), plus the Story Bank row id if a founders angle was used. - Spine path:
spine_idset to the spine's id,idea_idleft empty.sourcespopulated from whatever Story Bank row ids actually grounded the draft (the spine's own cross-references, and/or a founders-angle Story Bank row) — never a research note, since none exists for this path.
Leave viral_score at 0 and visual_ids empty — scoring the draft and
generating visuals are later phases (Quality/Critic Agent, Visual Agent),
not this skill's job.
10. Update the idea (idea path only)
Flip the Idea Note's status: selected → drafted. Spine path: no status
flip — Story Bank rows aren't a status-driven lifecycle, so the spine stays
as-is and can be reused or referenced again later.
Report back
Show the full draft text, hashtags, character count, and:
- Which research sources (idea path) or Story Bank row(s) (spine path) backed it.
- Which hook formula was used, and why — the engagement-goal match and the anti-repetition check against recent drafts.
- Whether a founders angle was attempted, and its outcome: used, with the Story Bank row it cited — or explicitly skipped because no real fit existed (never omit this without saying so).
This is pre-approval content, not ready to schedule. Remind the user that approval/edit/regenerate workflow (Phase 7) doesn't exist yet, so right now this just produces a file for manual review.
Hard rules
- Never invent a fact, statistic, or quote not present in the linked research notes (idea path) or the cited Story Bank rows (spine path). If the grounding doesn't support a specific claim the hook implies, soften the claim rather than fabricate support.
- Never draft an idea that isn't
status: selected(run/plan-weekfirst). - Never invent a founder anecdote or number that isn't present in
Content-Learnings/story-bank.md— no genuine match means no founders angle for this draft, not a fabricated one. - Never select a
proposed-status hook formula or founders angle for drafting — both must becanonical. - A hook formula shapes phrasing only; it must never misrepresent certainty or blend a personal opinion into the same declarative register as a sourced fact (REQUIREMENTS.md §21).
- The
--spinepath is grounded in Story Bank citations with the same rigor the idea path applies to research citations — never looser, never treated as "softer" evidence just because it's personal material. - Never mark anything as approved, scheduled, or published — this skill only produces a draft file.
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
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.

