priyanshu-aryapriyanshu-arya/LinkedIn_Automation/.claude/skills/critique-draft-substack-article/SKILL.mdcritique-draft-substack-article
Skill0 starsChanged 58 days agoMITView on GitHub
Use when the user asks to review/critique a Substack article's quality, check it for accuracy or duplication, compute its viral potential score, or explicitly invokes /critique-draft-substack-article. Runs the Critic Agent on a status:draft substack-article note, auto-revises it once if it scores low, and hands it off to approval by moving it to status:in_review.
In this repository38 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- audit-draftSkill
- check-plagiarismSkill
- critique-draftSkill
- 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-draftSkill
- write-draft-substack-articleSkill
- write-draft-substack-noteSkill
- write-draft-xSkill
The file
We split this skill into the rules it gives an agent
4 rules. About 0 are common, 2 are distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 76
Never invent a "historical performance" trend with no data — score neutral and say why.
- Distinctive: line 78
Never revise more than once per run.
- Very specific: line 79
Never move a genuine duplicate to in_review.
- Very specific: line 80
Never advance to anything beyond in_review — approval is human-only.
≈ “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
Critique Draft Substack Article (Critic Agent — Substack Articles)
Same process shape as /critique-draft (re-verify, dedup-check, score,
auto-revise once, hand off) but a genuinely different rubric — an article
is judged on depth and structure, not a scroll-stopping hook.
Arguments
- draft-id (optional) — critique one specific article. If omitted, run
against every
type: substack-articlenote currentlystatus: draft.
Process (per draft)
1. Re-verify accuracy
Re-read sources[]. Confirm every claim in the article body still traces
to the linked research. Fix drift now — soften or remove, don't just flag.
2. Originality / duplicate check
Scan Content-Learnings/content-index.md filtered to platform:
substack-article for the same underlying story/argument/structure as an
existing article. Same severity split as /critique-draft: a genuine
duplicate stops here (stay status: draft, append duplicate_flagged,
report plainly); a similar-but-distinct angle just pulls the Depth/
Originality sub-score down and still reaches human review.
3. Score Viral Potential (0-10)
Average these, each 0-10 — reweighted for long-form:
| Factor | How to assess it |
| Structure / flow | Do the sections actually build on each other, or could they be reordered with no loss? |
| Depth | Does this go meaningfully beyond what a feed post on the same topic would say? |
| SEO / discoverability fit | Do title, subtitle, seo_description, seo_tags actually reflect the content and a real search intent? |
| Technical accuracy | Consistency with linked research |
| Originality | Inverse of step 2's findings |
| Reader payoff | Does the closing land an actual point of view, not a restated summary? |
| Historical playbook alignment | Check Content-Learnings/playbook-substack.md's Articles section for a matching evidenced rule; otherwise neutral (5), say plainly no data exists yet |
Note there is no "hook strength" factor here in the feed-post sense — the title/subtitle/opening paragraph are judged under Structure/Reader payoff instead, since Substack's context (a subscriber who already opted in) is different from a scroll-stopping feed hook.
4. Auto-revise once if weak
Same rule as /critique-draft: below 6, one targeted revision pass
re-reading voice-guide-substack.md's Articles section, recompute, stop
after one pass.
5. Update the note
Same field updates as /critique-draft (viral_score, history,
## Critic Notes in the body, status: in_review) — skip entirely if
step 2 found a genuine duplicate.
Report back
Final score, per-factor breakdown, whether auto-revised, any similar-but-
distinct flags, and word count. A note that reached in_review is waiting
for human approval — nothing has been approved or published.
Notifications
Same batching rule as /critique-draft: one notification per run.
Hard rules
- Never invent a "historical performance" trend with no data — score neutral and say why.
- Never revise more than once per run.
- Never move a genuine duplicate to
in_review. - Never advance to anything beyond
in_review— approval is human-only.
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.