agentleFS
priyanshu-aryapriyanshu-arya/LinkedIn_Automation/.claude/skills/critique-draft-x/SKILL.md

critique-draft-x

Skill0 starsChanged 2 months agoMITView on GitHub

Use when the user asks to review/critique an X draft's quality, check it for accuracy or duplication, compute its viral potential score, or explicitly invokes /critique-draft-x. Runs the Critic Agent on a status:draft, platform:x Draft Note, auto-revises it once if it scores low, and hands it off to approval by moving it to status:in_review.

The file

Files
SKILL.md

Critique Draft X (Critic Agent — X pipeline)

X's sibling to /critique-draft. Same process shape and same evidence-gated rules — re-verify, dedup-check, score, auto-revise once if weak, hand off to human review. Never approves, schedules, or publishes.

Arguments

  • draft-id (optional) — critique one specific X draft. If omitted, run against every platform: x Draft Note currently status: draft.

Process (per draft)

1. Re-verify accuracy

Same as /critique-draft: re-read sources[], confirm every claim still traces to the linked research. Fix drift now, don't just flag it.

2. Originality / duplicate check

Scan Content-Learnings/content-index.md filtered to platform: x rows (plus a spot-check against LinkedIn/Substack rows for the same underlying story, since publishing the identical angle same-week across platforms without differentiation is still worth flagging, even if it's not a same-platform duplicate). Same severity split as /critique-draft:

  • Genuine duplicate (same story/example/hook/conclusion as an existing X post) — stop here, don't score, keep status: draft, append a duplicate_flagged history entry, report plainly.
  • Merely similar but distinct — note it, let it pull the Originality sub-score down, still goes to human review.

3. Score Viral Potential (0-10)

Average these, each 0-10 — same shape as /critique-draft with one added factor for threads:

FactorHow to assess it
Hook strengthDoes tweet 1 stand alone and stop the scroll?
ScannabilityShort, punchy, no wall-of-text tweet in the sequence
Novelty / insightFresh beyond basic regurgitation
Discussion catalysisWould this get real replies, not just likes?
Technical accuracyConsistency with linked research
Thread cohesion (threads only; score 10 for a single post)Does each tweet stand alone reasonably well while the sequence still reads as one throughline, no filler tweets just to extend the thread?
OriginalityInverse of step 2's findings
Historical playbook alignmentCheck Content-Learnings/playbook-x.md for a matching evidenced rule; otherwise neutral (5), say plainly no data exists yet

4. Auto-revise once if weak

Same rule as /critique-draft: below 6, exactly one targeted revision pass re-reading voice-guide-x.md, recompute, stop after one pass regardless.

5. Update the Draft Note

Same field updates as /critique-draft (viral_score, history, ## Critic Notes, status: in_review) — skip entirely if step 2 found a genuine duplicate.

Report back

Same shape as /critique-draft's report, plus whether this draft is a single post or a thread and its tweet count.

Notifications

Same batching rule as /critique-draft: one batched notification per run covering however many X drafts reached in_review.

Hard rules

Identical to /critique-draft's hard rules, platform-scoped to X drafts.

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.