Menu
Use when you need deep codebase exploration to find code, patterns, or references.
Installer avec Codex ou Claude Copiez ce prompt, collez-le dans Codex, Claude ou un autre assistant, puis laissez-le vérifier la page du skill et l'installer pour vous.
Basé sur la classification professionnelle SOC
| name | team-shinchan:deepsearch |
| description | Use when you need deep codebase exploration to find code, patterns, or references. |
| user-invocable | false |
If args is empty or only whitespace:
Ask user: "What would you like to search for?"
STOP and wait for user response
If args length > 2000 characters:
Truncate to 2000 characters
Warn user: "Request was truncated to 2000 characters"
Do not read further. Execute these Tasks NOW:
// Step 1: Fast Search (Shiro)
Task(
subagent_type="team-shinchan:shiro",
model="haiku",
prompt=`/team-shinchan:deepsearch has been invoked.
## Codebase Exploration Request
Perform fast search:
- File name pattern matching
- Keyword search
- Directory structure analysis
Search target: ${args || '(What to search)'}
`
)
// Step 2: Deep Search if needed (Masumi)
// Call additionally if Shiro results are insufficient
Task(
subagent_type="team-shinchan:masumi",
model="sonnet",
prompt=`Perform deep analysis based on Shiro search results:
- Code content analysis
- Related documentation search
- Dependency tracking
Search target: ${args || '(What to search)'}
`
)
STOP HERE. The above Tasks handle everything.
Use when you have a large-scale, multi-phase project requiring orchestrated execution.
Use when you need persistent looping until a task is fully complete.
Deterministic adversarial code review for high-stakes scope — independent per-dimension review, a non-skippable per-finding refutation, completeness + interaction critics, and a deterministic 3-lens rubric judge panel. Opt-in main-loop Workflow tier.
Use when the user wants to review accumulated skill feedback, verdict trends, or improvement candidates collected during Stage 4 retrospectives. Trigger on "show skill feedback", "스킬 피드백 보여줘", or finding which skills need /writing-skills work.
Deterministic competitive code tournament — N builders independently solve one task and return patches, an Action-Kamen judge scores them head-to-head, the winner is picked by score and applied. Opt-in main-loop Workflow tier.
Deterministic adversarial debate for high-stakes or irreversible decisions — mandatory refutation plus a scored judge panel. Opt-in main-loop Workflow tier.