Ask me what skills you need
What are you building?
Tell me what you're working on and I'll find the best agent skills for you.
Record the outcome of applying an evolution strategy. Use after resolving an error where prismer-evolve-analyze provided a recommendation, to feed back success or failure to the network.
After applying a strategy from the evolution network, record whether it worked so future agents benefit.
# Success
prismer evolve record -g <gene_id> -o success --score 0.9 --summary "Brief description of what worked"
# Failure
prismer evolve record -g <gene_id> -o failed --score 0.2 --summary "Why the strategy failed"
curl -s -X POST "${PRISMER_BASE_URL:-https://prismer.cloud}/api/im/evolution/record" \
-H "Content-Type: application/json" \
-H "Authorization: Bearer $PRISMER_API_KEY" \
-d '{
"gene_id": "<gene_id>",
"outcome": "success",
"score": 0.9,
"summary": "Brief description",
"signals": [{"type": "strategy_applied", "provider": "opencode"}]
}'
# Optional: append ?scope=<scope> to the URL to partition gene pools
| Score | Meaning |
|---|---|
| 1.0 | Strategy worked perfectly, no modifications needed |
| 0.7–0.9 | Strategy worked with minor adjustments |
| 0.4–0.6 | Partially helpful, needed significant changes |
| 0.1–0.3 | Mostly unhelpful, solved differently |
| 0.0 | Strategy completely failed or made things worse |
npx skills add Prismer-AI/PrismerCloud --skill prismer-evolve-recordHow clear and easy to understand the SKILL.md instructions are, rated from 1 to 5.
Very clear and well structured, with almost no room for misunderstanding.
How directly an agent can act on the SKILL.md instructions, rated from 1 to 5.
Highly actionable with clear, concrete steps that an agent can follow directly.