Use when adding new error messages to React, or seeing "unknown error code" warnings.
npx skills add front-depiction/claude-setup --skill "parallel-explore"
Install specific skill from multi-skill repository
# Description
Parallel exploration of codebase questions by decomposing into independent tracks. Use when exploring architecture, understanding systems, or investigating complex questions that benefit from multiple perspectives.
# SKILL.md
name: parallel-explore
description: Parallel exploration of codebase questions by decomposing into independent tracks. Use when exploring architecture, understanding systems, or investigating complex questions that benefit from multiple perspectives.
exploration := decompose(question) β parallel(tracks) β aggregate(findings)
track := independent β§ parallelizable β§ logical-unit
agent := explorer(track) β findings(structured)
synthesis := merge(findings) β coherent(answer)
analyze :: Question β [Track]
analyze question = do
dimensions β identify (orthogonal (aspects question))
granularity β ensure (|dimensions| >= 3 β§ |dimensions| <= 6)
tracks β map toTrack dimensions
verify (independent tracks)
pure tracks
-- each track explores one dimension without overlap
-- tracks are logical units of investigation
-- minimum 3 tracks to ensure sufficient coverage
data Track = Track
{ name :: String -- short identifier
, focus :: Scope -- what to investigate
, approach :: Strategy -- how to investigate
, gates :: [Gate] -- verification checkpoints
, deliverable :: Artifact -- what to return
}
data Gate = Gate
{ checkpoint :: Condition -- what must be true
, verification :: Command -- how to verify
}
-- each track subdivides into gated steps
-- step N must pass gate before step N+1
trigger :: Question β Bool
trigger question
| complex (architecture question) = True
| multiple (dimensions question) = True
| benefits (parallelPerspectives) = True
| investigating (multiCauseBug) = True
| researching (crossCuttingPatterns) = True
| otherwise = False
decompose :: Question β Effect [Track]
decompose question = do
aspects β identify (dimensions question)
tracks β traverse toTrack aspects
verify (all independent tracks)
pure tracks
dispatch :: [Track] β Effect [Agent]
dispatch tracks = parallel $ fmap spawnExplorer tracks
spawnExplorer :: Track β Effect Agent
spawnExplorer track = spawn Explore (prompt track)
aggregate :: [TrackFindings] β Synthesis
aggregate findings = do
common β intersect (conclusions findings)
divergent β difference (conclusions findings)
gaps β union (uncertainties findings)
confidence β assess divergent
pure $ Synthesis { common, divergent, gaps, confidence }
-- ADDITION to parallel tracks, not replacement
-- When complex modules detected β spawn module specialists in parallel with tracks
-- Targets BOTH Effect library modules AND internal codebase modules
detect :: Code β [Module]
detect code = filter isComplex (modulesUsed code)
isComplex :: Module β Bool
isComplex m = isEffectModule m β¨ isCodebaseModule m
isEffectModule :: Module β Bool
isEffectModule m = m β { Stream, Layer, Fiber, Scope, Schema, Effect.gen, Match }
isCodebaseModule :: Module β Bool
isCodebaseModule m = m β /modules -- internal modules with ai-context.md
data ModuleSource = EffectLibrary | CodebaseInternal
moduleSource :: Module β ModuleSource
moduleSource m
| isEffectModule m = EffectLibrary
| isCodebaseModule m = CodebaseInternal
spawn :: Module β Effect Specialist
spawn m = do
context β case moduleSource m of
EffectLibrary β /module m -- external Effect documentation
CodebaseInternal β /module path -- internal ai-context.md
scope β extractUsage m code -- ONLY how m is used
pure $ Specialist
{ context := context -- module documentation only
, scope := scope -- usage patterns in code
, goal := efficiency -- missed conveniences, anti-patterns
, constraint := reviewOnly -- no implementation, just review
}
You are a [MODULE] specialist.
Your ONLY context: the /module [MODULE] documentation
Your ONLY scope: how [MODULE] is used in the provided code
Your ONLY goal: identify efficiency issues
Review for:
- Missed conveniences (simpler APIs available)
- Anti-patterns (common misuse)
- Idiomatic alternatives (more Effect-like approaches)
Output:
- Specific file:line citations
- What could be improved
- How to improve it (idiomatic pattern)
Do NOT:
- Implement changes
- Review anything outside [MODULE] usage
- Consider broader architecture
explore :: Question β Code β Effect Synthesis
explore question code = do
-- Regular parallel tracks (existing behavior)
tracks β decompose question
trackAgents β dispatch tracks
-- ADDITION: module specialists (orthogonal overlay)
modules β detect code
specialists β parallel (spawn <$> modules)
-- Aggregate both
trackFindings β await trackAgents
specialistFindings β await specialists
aggregate (trackFindings ++ specialistFindings)
execute :: Track β Effect Findings
execute track = do
context β gatherContext track
gateβ β verify (understood context)
findings β explore context track
gateβ β verify (evidence findings)
summary β synthesize findings
gateβ β verify (complete summary)
pure summary
gate :: Checkpoint β Effect ()
gate = \case
Understood ctx β can describe files/patterns found
Evidence fnd β every claim has file:line citation
Complete sum β summary answers track question
deepen :: Track β Effect Synthesis
deepen track = do
subtracks β decompose (question track)
agents β dispatch subtracks
findings β await agents
aggregate findings
-- apply same decomposition recursively for deeper exploration
# Supported AI Coding Agents
This skill is compatible with the SKILL.md standard and works with all major AI coding agents:
Learn more about the SKILL.md standard and how to use these skills with your preferred AI coding agent.