Sunwood-ai-labs/SourceSage

Documents & files

Graded A on context cost and craft. 112 tokens of this skill ride in the system prompt of every message, whether or not it fires.

Ranked #12 of 43 on the leaderboard →

Score: 3.70 / 4
Resident: 112 tokens
Footprint: light
Per activation: 885 tokens

Share this grade on X ↗Re-grade now →View on GitHub ↗Grade another skillLast graded Sep 30, 2026 from HEAD

Add the badge to your README

It shows the current letter grade and links back to this page. Improve the skill, hit Re-grade now, and the badge follows — GitHub may keep the old image for up to an hour.

Markdown

[![Skill Grader: A](https://seoagent.com/skill-grader/badge/Sunwood-ai-labs/SourceSage.svg)](https://seoagent.com/skill-grader/Sunwood-ai-labs/SourceSage)

HTML

<a href="https://seoagent.com/skill-grader/Sunwood-ai-labs/SourceSage"><img src="https://seoagent.com/skill-grader/badge/Sunwood-ai-labs/SourceSage.svg" alt="Skill Grader: A" height="20"></a>

Report card

A overall · 3.70/4

github.com/Sunwood-ai-labs/SourceSage@HEAD · 1 skill · 112t resident (light) · 885t per activation · 15,219t bundle

Resident footprint

112t of description text ride in the system prompt of every message (paper's observed per-skill range: 50–280t).

Description honesty

Reads as a statement of purpose, not defensive trigger padding.

Body size

Largest body is 525 words (~885t), loaded on every trigger. Corpus median 921, p90 2207.

Progressive disclosure

28 auxiliary .md files ship but the body never points at them — the agent won't find them.

Factoring

Scoped to a coherent procedure set.

CLI leverage

Drives a CLI (`uv`, 6 invocations) — verification and state live in code, not prompt.

A overall (3.70/4). 1 skill(s), 112t resident on every message (light), 885t loaded per activation. Weakest axis: progressive disclosure (C).

Six deterministic axes, no LLM, same result every run. Tokens are approximated as chars/4. Read the methodology, or see how SEO skills score on task outcomes in the open SEO skill benchmark.