Ready to supercharge your AI agents?

Install any skill in seconds. Free to start, no credit card required.

Search And Plan · optim-agent | DecimalAI

researchby optim-agent

Search And Plan

optim-agent/search-and-plan

Source optim-agent/optim-plans ↗License MIT

Use when external research should inform a repo-change plan before implementation; exclude direct implementation-only, factual/explanation, trivial, or explicit no-plan requests.

SkillScore45Maintained

Automatic

Project Management pack

Last verified 8/24/2026

What it changes

Passed 5 of 10 test cases with the skill, and 2 of 10 without it.

Without skill23%

With skill45%

Checks it passed that the model alone missed

Restricts the high-signal source pack to between 3 and 7 sources.
Continues the research workflow using repository evidence as fallback.
Records candidate choices using controller choice prompts before plan drafting.

and 4 more in the full report

7 fixed3 passed either way10 still failing2 broke

Pass-rate lift: +23 pp
Tokens: -23%
Agent turns: 0%
Response time: -57%

Model gemini-3.6-flashMeasured 8/24/2026v110 of 22 cases passed with skill · 10 comparable

Tokens, turns and time describe resource use, not a dollar price.

Use this skill

Save it into your project's skills folder. Claude Code picks it up in its next session.

curl -fsSL --create-dirs -o .claude/skills/optim-agent-search-and-plan/SKILL.md https://app.decimal.ai/s/optim-agent-search-and-plan/SKILL.md

Want it pinned and updatable? pip install "decimalai>=0.15", then decimalai skills add optim-agent-search-and-plan, and start Claude Code with claude --plugin-dir "$(git rev-parse --show-toplevel)/.agents"

1 · One-time — install the CLI (docs ↗)

pip install "decimalai>=0.15"

2 · Get the skill — writes .agents/skills/optim-agent-search-and-plan/ and pins it in .decimal/skills.lock

decimalai skills add optim-agent-search-and-plan

No CLI? Get the same file — curl · raw file

curl -fsSL --create-dirs -o .claude/skills/optim-agent-search-and-plan/SKILL.md https://app.decimal.ai/s/optim-agent-search-and-plan/SKILL.md

Raw file: /s/optim-agent-search-and-plan/SKILL.md ↗ · machine-readable summary at /s/optim-agent-search-and-plan.json ↗

Safety checks: 3 of 3 passed

Code scan CleanIntent PassedContent Clean

See it in action

One of 7 test cases the skill fixed. Same model, same prompt, with and without the skill.

1 of 7

Prompt

You found 12 blog posts and 4 documentation pages about React Server Components performance. How many sources should be included in the source pack?

What was checked, without and with the skill
Checked	Without skill	With skill
Restricts the high-signal source pack to between 3 and 7 sources.	Missed	Passed

case-15 · gemini-3.6-flash · 8/24/2026 · answers are capped and redacted

What's inside

The instructions your model reads when the skill loads.

SKILL.md35 linessingle file

Search And Plan

Research first; planning and execution still follow ../optim-plans/SKILL.md.

Required Flow

Read and follow ../optim-plans/SKILL.md, preserving explicit mini-plan, small-plan, plan, big-plan, and huge-plan routing.
Inspect the target Git repo read-only before asking product questions.
Initialize or resume the optim-plans controller before external research.
Perform read-only initial research. Prefer agent-reach when available; if missing, give one sentence of install guidance, do not install agent-reach or any other tool before execution approval, and continue with fallback search/repo evidence.
Ask the first evidence-informed question as a controller-backed optim-plans choice prompt, and submit/record that answer through the controller before writing refs.
Only after that recorded controller answer, write sources and REF_ANALYSIS.md under docs/optim-plans/YYYY-MM-DD-topic/refs/search-and-plan/<topic>/.
Ask any later product questions only after refs are persisted.
Continue into the selected optim-plans level: PLAN_v1.md, refinement, immutable execution approval, execution, validation, controller verification, and integration gates.

Research Contract

Keep a 3-7 high-signal source pack; if fewer than 3 credible sources exist, record the gap instead of padding with weak sources.
Record attempted queries, attempted backends, backend failures with reasons, why sources were insufficient, and then continue from repository evidence.
Backend failure fallback is valid only when the failure reason is recorded.
Do not write pre-execution ./refs/ files; refs live only under the run artifact path above.
Do not require a strict source manifest before planning.

REF_ANALYSIS.md

Use these sections exactly:

## Sources
## Findings
## Adoptable ideas
## Risks/not-applicable points
## Evidence gaps
## Candidate user decisions

Every adoptable idea must be presented as an evidence-backed optim-plans choice prompt and recorded through the controller before it can be included in any plan.

Track record

Benchmark

5 / 10

test cases passed, vs 2 / 10 without

Adoption

Users: 0
Times used: 0

Safety and source details

SkillSafety

✓ 3 of 3 passed

Code scan15 of 15 checks clean

0 critical · 0 warning · 0 info — scanner v4 — 8/3/2026

View 15 scanner checks

live_secretremote_code_executionreverse_shelldata_exfiltrationssrf_metadataagent_snoopinganti_refusaldestructive_commandsobfuscated_unicodeunicode_homoglyphopaque_blobsuspicious_urltool_overreachtrigger_abuseprompt_injection_phrasing

IntentSafe

LLM judge · does the skill do what it claims

ContentClean

LLM judge · policy & harmful-content review

Source

Imported

optim-agent / optim-plans

skills/search-and-plan/SKILL.md

View on GitHub ↗

License MITImported 8/3/2026Version v1

Other skills for the same job

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.

Ready to supercharge your AI agents?

See it in action

What's inside

Search And Plan

Required Flow

Research Contract

REF_ANALYSIS.md

Track record

Benchmark

Adoption

SkillSafety

Source

Other skills for the same job

Related skills