Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →optim-agent/reference-before-plan
Research references before creating a Prime Agent plan.
Passed 7 of 12 test cases with the skill, and 4 of 12 without it.
and 6 more in the full report
8 fixed4 passed either way6 still failing4 broke
Tokens, turns and time describe resource use, not a dollar price.
One of 7 test cases the skill fixed. Same model, same prompt, with and without the skill.
We finished reading a technical blog post on zero-downtime database migrations. Which command logs the structured analysis artifact, and what gets logged if a structured parser is not available?
| Checked | Without skill | With skill |
|---|---|---|
| Uses record-reference-analysis to record structured analysis | Missed | Passed |
| Uses record-reference-waiver if a manual structured read is performed due to lack of a richer parser | Missed | Passed |
The instructions your model reads when the skill loads.
Use this skill when external references must shape the plan before implementation.
../_shared/references/prime-planning-workflow.md, ../_shared/references/state-and-config.md, and ../_shared/references/plan-artifact-template.md.python3 scripts/optim_plans.py init --repo <target-repo> --topic <topic> --plan-level reference-before-plan --request-text <original-request>..git/optim-plans for controller state and docs/optim-plans/YYYY-MM-DD-topic/ for public artifacts.PLAN_v1.md. Store large downloads outside the target repo by default under ~/.cache/optim-plans/refs/<repo-slug>/<topic>/.record-reference: title, URL, kind, retrieval method, access date, local path, coverage, and evidence gaps.record-reference-analysis. If a richer parser is unavailable, do a manual structured read and record any waiver with record-reference-waiver.PLAN_v1.md. Each question must be based on downloaded content, put the recommended option first, put Other second-last, and put Auto-complete last.Auto-complete cannot grant this waiver.Include original request, repo evidence, attempted queries, selection criteria, selected and rejected references, local paths, structured analysis paths, adoption questions and answers, accepted ideas, rejected ideas, evidence gaps, user-granted waivers, and reviewer/criticizer settings used for the run.
Reference ideas are not eligible for PLAN_v1.md until their adoption question answers are recorded.
test cases passed, vs 4 / 12 without
Published skills whose descriptions closely match this one, with their headline benchmark lift.
More skills from the same source and nearby in the registry, with their headline benchmark lift.