lyndonkl/claude/compute-baseline

Trust tier 2
  • community
LLM review pending
Source check pending

Checking that the SKILL.md file is still in the repository.

Computes substacker's rolling 4-week baseline for open rate, click rate, views-per-send, and weekly subscriber delta using corpus/stats/ archived CSVs. Produces per-metric z-scores of the current week against the baseline and flags cold-start windows where fewer than 4 prior weeks exist. Use after ingest-substack-csv each Monday. Trigger keywords — baseline, rolling median, z-score, cold start, per-metric comparison.

Install

npx vskill@latest install lyndonkl/claude/compute-baseline

SKILL.md

Compute Baseline

Workflow

Per week:
- [ ] Step 1: Load last 4 weekly CSVs from corpus/stats/
- [ ] Step 2: For each metric (open_rate, click_rate, views_per_send, weekly_sub_delta):
    - Compute 4-week median
    - Compute trimmed median (drop 1 outlier)
    - Compute IQR
- [ ] Step 3: z-score = (current - median) / IQR
- [ ] Step 4: If <4 weeks in history, return baseline: not-yet-established
- [ ] Step 5: Emit baseline object per metric with confidence flag

Baseline object schema

json
{
  "open_rate": {"current": 0.47, "median_4w": 0.49, "trimmed_median": 0.49, "iqr": 0.03, "z": -0.67, "confidence": "medium"},
  "click_rate": {...},
  "views_per_send": {...},
  "weekly_sub_delta": {...},
  "cold_start": false
}

Confidence: high if 4+ weeks and low IQR. medium if 4+ weeks and typical IQR. low if N<4.

Guardrails

  1. Only compare writer's own trajectory. Never pull external benchmarks.
  2. Below N=4, return cold-start flag. Report writes "baseline not yet established" rather than noisy averages.
  3. IQR-based comparisons (not σ-based) for robustness at small N.
  4. Outlier handling: use trimmed median to blunt a single extreme week.
  5. |z| ≥ 1.0 is the "material move" threshold used downstream by attribute-performance.
  6. Don't recompute baselines for prior weeks. Each week's baseline is from that week's trailing 4.

Security

  • Security scanPassed, 100 of 100Pattern checks on the skill's files.
  • Model reviewNot runA model review is scheduled for the next rescan.

See the full security report

Quality evaluation

  • Task evaluationNot runNo evaluation yet. Run one in Skill Studio.

See evaluation results

Source

Popularity

Repository stars
147
Repository forks
22
7-day trend score
25

Stars and forks are for the whole GitHub repository, not this skill alone. Not refreshed yet.

Works with

Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.

  • Claude Code
  • Cursor
  • GitHub Copilot
  • Windsurf
  • Codex
  • Gemini CLI
  • Cline
  • Amp
  • Roo Code
  • Goose
  • Continue
  • Kilo Code
  • Junie
  • Trae
  • OpenHands
  • Qwen Code
  • AI Agent
  • Tabnine Chat
  • Sourcegraph Cody
  • Amazon Q
  • Firebase Genkit
  • Sweep
  • Devin
  • Phind
  • Mentat
  • Claude Desktop
  • Aider
  • GPT Pilot
  • Plandex
  • bolt.new
  • v0
  • Replit Agent
  • Anse
  • TypingMind
  • BetterChatGPT
  • LibreChat
  • Warp AI
  • Pieces
  • Fig AI
Embed badge
vskill VERIFIED badge
[![vskill VERIFIED](https://verified-skill.com/api/v1/skills/lyndonkl/claude/compute-baseline/badge)](https://verified-skill.com/skills/lyndonkl/claude/compute-baseline)

Something wrong with this skill or its page? Tell us.