
Shinka Inspect
After you've run a Shinka evolution batch and have a SQLite results database sitting there, this pulls out your top performers and packages them into a single Markdown file you can hand to a coding agent. It ranks by combined_score, grabs the top k correct programs (or just top k if nothing passed), and bundles metadata, feedback, and code snippets together. The main value is skipping the manual database spelunking step between runs. You point it at your results directory, get a compact context file back, and use that to plan your next mutation round. Read-only on your artifacts, writes one markdown bundle, that's it.
- skills.sh installs
- 242
- GitHub stars
- 1.5K
- last updated
- Aug 21, 2026
Install
$ npx -y skills add sakanaai/shinkaevolve --skill shinka-inspect --agent claude-codeInstalls into .claude/skills of the current project.
Files
Shinka Inspect Skill
Extract the strongest programs from a Shinka run and package them into a context file that coding agents can load directly.
When to Use
Use this skill when:
- A run already produced a results directory and SQLite database
- You want to inspect top-performing programs before launching the next batch
- You want a compact context artifact instead of manually browsing the DB
Do not use this skill when:
- You still need to scaffold a task (
shinka-setup) - You need to run evolution batches (
shinka-run)
What it does
- Uses
shinka.utils.load_programs_to_dfto read program records - Ranks programs by
combined_score - Selects top-
kcorrect programs (fallback to top-koverall if no correct rows) - Writes one Markdown bundle with metadata, ranking table, feedback, and code snippets
Workflow
- Confirm run artifacts exist
ls -la <results_dir>
- Generate context bundle
python skills/shinka-inspect/scripts/inspect_best_programs.py \
--results-dir <results_dir> \
--k 5
Sponsors
Details
- Repository
- github.com/sakanaai/shinkaevolve
- Categories
- AI & Agent Building, Productivity & Planning
- First seen
- Jun 3, 2026