Install any skill in seconds. Free to start, no credit card required.
Get Started Free →notque/plant-seed
Capture forward-looking idea as a seed for future feature design.
Passed 14 of 18 test cases with the skill, and 7 of 18 without it.
and 7 more in the full report
9 fixed8 passed either way5 still failing0 broke
Tokens, turns and time describe resource use, not a dollar price.
One of 9 test cases the skill fixed. Same model, same prompt, with and without the skill.
When writing a deferred seed to disk during the write phase, where in the filesystem hierarchy is the primary index file written, and what directory contains archived items?
| Checked | Without skill | With skill |
|---|---|---|
| Identifies .seeds/index.json as the index storage file path. | Missed | Passed |
| Identifies .seeds/archived/ as the path for archived seeds. | Missed | Passed |
The instructions your model reads when the skill loads.
Capture forward-looking ideas with trigger conditions so they resurface at the right time. Seeds carry WHY (rationale) and WHEN (trigger), making them far more valuable than bare TODO comments.
Seeds are stored locally in .seeds/ (gitignored) and automatically surfaced during feature-lifecycle design phase when their trigger conditions match. This workflow is designed for deferred ideas only — if the user describes work that should happen now, suggest creating a task or issue instead.
Goal: Gather the idea, trigger condition, scope, and rationale from the user.
Key Constraint: This skill captures deferred ideas only. If the user describes something that should happen in the current session, suggest creating a task or issue instead. Planting a seed for immediate work means it never gets surfaced — it just gets forgotten in a different way.
Step 1: Understand the idea
Extract from the user's description:
If the user provides all three, proceed. If any are missing, ask:
For missing trigger condition:
When should this idea resurface? Describe the condition, e.g.:
- "when we add user accounts"
- "when performance optimization is needed"
- "when the API exceeds 10 endpoints"
For missing scope:
How big is this work? Small (< 1 hour), Medium (1-4 hours), or Large (4+ hours)?
For missing rationale:
Why does this matter? What's the insight behind this idea? (e.g., "Response times degrade linearly with DB calls per request -- at 10+ endpoints the shared query pattern becomes the bottleneck")
Step 2: Generate seed ID
test cases passed, vs 7 / 18 without
More skills from the same source and nearby in the registry, with their headline benchmark lift.