Skip to main content

/sentinel

Source: ~/.claude/skills/sentinel/SKILL.md

name: sentinel version: 2.0 description: > Run full system audit using 5-agent team (BA, Architect, Developer, Tester, Validator). Use when: "audit the system", "run sentinel", "system health check", "/sentinel", "review infrastructure", "find issues", "what's broken". argument-hint: "[target] — e.g. 'tools', 'hooks', 'Drop project', 'daemons', or empty for full audit" level: 4 company: ALAI

/sentinel — System Audit Team

DescriptionPurpose

Spawns a 5-agent audit team (BA, Architect, Developer, Tester, Validator) that performs a comprehensive systemparallel audit andthat delivers a consolidated report with prioritized action items. BA + Architect + Developer + Tester run in parallel → Validator consolidates.

InvocationVariables












[target]
  • No target = 
    • audit
  • Target
    • examples:"tools","hooks","Dropproject","daemons"






    VariableTypeDescriptionDefault
    /sentineltarget
string
Audit 
      scope
    		
full system

    

    
model
string
Agent "offline",model
sonnet

    

    
depthstringshallow | deepdeep
    

    Team Composition
    

    


    
Role
Agent File
Focus

    
    


    
BA
sentinel-ba.md
Business value, gaps, redundancy, ROI

    

    
Architect
sentinel-architect.md
Architecture, integrations, offline/online parity

    

    
Developer
sentinel-developer.md
Code quality, dead code, tech debt, bugs

    

    
Tester
sentinel-tester.md
Functional testing, daemon health, data integrity

    

    
Validator
sentinel-validator.md
Cross-reference, consolidate, final action plan

    
    

    

    Execution FlowWorkflow
    

    Phase 1: Pre-flight
    

      

    • Read audit target/scope from $ARGUMENTS
      • 

    • if no target → set target = "full system"
      • 

    • if target = "quick" → set depth = shallow (skip code quality, focus on daemons + health)
      • 

    

    Phase 2: Parallel Audit (4 agents)agents simultaneously)
    

    Spawn BA,4 Architect, Developer, and Tester as parallel Task sub-agents (sonnetin model).parallel, Eacheach agent receives:with:
    

      

    1. Their role-Role-specific prompt from ~/.claude/agents/sentinel-{role}.md
      2. 

    2. TheAudit audit target/scopetarget
      3. 

    3. Key systempaths: paths~/system/, to~/.claude/, investigate~/system/databases/
      4. 

    

    Task(subagent_type="general-purpose", model="sonnet", name="sentinel-ba")[Parallel]:
  Task(subagent_type="general-purpose",sentinel-ba)        model="sonnet",→ name="sentinel-architect")business audit report
  Task(subagent_type="general-purpose",sentinel-architect) model="sonnet",→ name="sentinel-developer")architecture audit report
  Task(subagent_type="general-purpose",sentinel-developer) model="sonnet",→ name="code quality report
  Task(sentinel-tester")tester)    
    → 
    Phasehealth/functional 2: Validation (1 agent)
    

    After all 4 complete, spawn Validator with ALL 4 reports as input.
    

    Task(subagent_type="general-purpose", model="sonnet", name="sentinel-validator")report

    

    Phase 3: DeliveryValidation (after all 4 complete)
    

    Spawn Validator with all 4 reports as input:
    

    Task(sentinel-validator, input=[ba_report, arch_report, dev_report, test_report])
  → consolidated final report

    

    Phase 4: Output
    

      

    • SavePrint final report tofrom ~/system/reports/sentinel-audit-{date}.mdValidator
      • 

    • Postif summarycritical toissues HiveMindfound → create MC tasks via delegate_task
      • 

    • Presentif findingsminor toissues user→ list as recommendations
      • 

    

    AgentReport Constraints
    

      

    • Model: sonnet (NEVER opus for agents)
      • 

    • All agents: READ-ONLY. No file modifications.
      • 

    • Tester: Only runs safe, non-destructive commands
      • 

    • Each agent produces structured markdown report
      • 

    

    Key System Paths for AgentsFormat
    

    ~/.claude/hooks/SENTINEL —AUDIT HookREPORT
scriptsTarget: [scope]
Date: [timestamp]
Model: [sonnet|opus]

CRITICAL (Python/Bash)fix ~/.claude/skills/immediately):
  —[numbered Skilllist]

definitionsHIGH ~/.claude/agents/(fix —this Agentweek):
  definitions[numbered ~/.claude/mcp.jsonlist]

—MEDIUM MCP(backlog):
  server[numbered configlist]

~/.claude/settings.jsonMC —Tasks ClaudeCreated: settings[list +of permissionstask ~/system/tools/IDs]
—Next AllAudit: system[recommended tools
~/system/tools/manifest.md — Tool registry
~/system/rules/           — Business & dev rules
~/system/agents/hivemind/ — Knowledge base
~/system/db/              — SQLite databases
~/system/config/          — Configuration files
~/system/context/         — Context documents
~/system/specs/           — Specifications
~/system/boot.sh          — Boot sequence
~/Library/LaunchAgents/   — Daemon definitionsinterval]

    

    Reuse
    

    This skill is designed to be called repeatedly. Each run produces a dated report.
Compare reports over time to track system health trajectory.$ARGUMENTS
    Back to top