agentleFS
Sign inStart free
ttalkkaklabttalkkaklab/social-flow/skills/review-recent/SKILL.md

review-recent

Skill1 starsChanged 3 months agoApache-2.0View on GitHub

Reports how the last few published videos did on YouTube and Instagram. Use when the user asks to "최근 영상 피드백", "유튜브 인스타 평가", "게시분 보고서", "지표 보고서 만들어", "review recent posts", "feedback on the last videos". Calls content_feedback for the last 5 posts per platform, writes a chart-first HTML report under data/[channel]/growth/review-recent.html, and opens it.

The file

We split this skill into the rules it gives an agent

4 rules. About 0 are common, 3 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    If both fail, stop and point at setup-youtube / setup-instagram.

    line 65
  2. Distinctive:

    Don't have a human re-polish the report.

    line 69
  3. Very specific:

    Don't change growth loop state (growth-plan.md).

    line 70
  4. Distinctive:

    Don't call publish tools.

    line 71

≈ “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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Automatic execution policy: Read automation guidance first. Creative rules and review findings guide improvements; scores, review/hash/freeze records, CTA choices and intermediate approval steps do not block authorized work. Preserve actual external-action authorization and stop at the artifact the user requested. This policy supersedes older quality-gate wording below.

Feedback on recent posts — HTML report

Scores the latest YouTube and Instagram posts in separate platform sections. Don't spell the scores out in chat. Write an HTML built around tables, funnels and bar charts, then open that file.

/social-flow:review-recent my-channel
/social-flow:review-recent my-channel 5

Procedure

  1. Confirm the channel — with no argument, look through data/*/profile.md and pick the channel. If there's no data/<slug>/profile.md, point at /social-flow:channel add and stop.
  2. Call the tool — mcp__social-flow__content_feedback
    • channel: slug
    • limit: the argument, or 5 (max 10)
    • days: 28
    • outputPath: data/<slug>/growth/review-recent.html (relative to cwd)
  3. Archive copy — copy it once more into the same folder as review-recent-YYYYMMDD-HHMM.html (overwrite history).
  4. Open it — if ego lite is around, open the report over file://. Otherwise just report the path.
  5. Report in chat — one line with the path + per platform "fix N · watch N · pending N". The per-episode numbers belong to the HTML. Don't redraw the table in chat.

Metrics (the tool is the source of truth)

The comparison baseline is the median across these N episodes. No absolute lines like YouTube's 70% at 30 seconds.

PlatformWhat to look atLever to pull when it drops
YouTubeearly pass-through engagedViews/views · average view % · views · shares against engaged viewshook · retention · angle (if views alone are low while hook and retention hold, treat the title as the problem) · shares
Instagram reels3-second drop-off · average watch seconds · shares against reachhook · retention · shares

Shares read against engagedViews on YouTube — the views that got past the opening — because that platform gives no reach number, and dividing by raw views would fold a weak opening into the share rate and bill one problem twice. Instagram reads them against reach. Same lever either way: a forward puts the video in front of someone the recommendation never chose.

YouTube click-through rate, subscribers per episode, and per-reel follows on Instagram are either not given by the platform or unstable. Leave those cells empty in the report rather than pretending they don't exist.

A platform without a token gets "none" in that section only; write the rest. If both fail, stop and point at setup-youtube / setup-instagram.

Rules

  • Don't have a human re-polish the report. The HTML the tool wrote is the artifact.
  • Don't change growth loop state (growth-plan.md). Observe only.
  • Don't call publish tools.

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.

Check your setup, free