/redteam Command

Adversarial red team analysis of the current conversation or specified target.

Usage

/redteam [mode] [target]

Arguments

mode (optional):

• quick - Fast 2-3 vector analysis, skip grounding
• standard - Balanced 5-6 vectors with basic grounding (default)
• deep - All categories + meta-analysis with full grounding
• focus:[category] - Deep dive on specific category (e.g., focus:reasoning-flaws)

target (optional):

• conversation - Current conversation context (default)
• file:path - Analyze specific file
• code - Analyze recent git changes

Instructions

You are the MINIMAL entry point for red team analysis. Your ONLY job is to:

1. Check PAL MCP availability (optional enhancement)
2. Parse the mode and target from arguments
3. Extract a structured context snapshot
4. Launch the red-team-coordinator agent with the snapshot
5. Return the coordinator's output directly to the user

Step 1: Check PAL Availability (Non-Blocking)

Launch the pal-availability-checker agent to detect if PAL MCP is available:

Task: Launch pal-availability-checker agent
Agent: agents/pal-availability-checker.md
Prompt: Check if PAL MCP is available and list models

Parse the YAML result and include pal_available: true/false in the snapshot. This step is NON-BLOCKING - continue regardless of result. PAL is optional.

Step 2: Parse Arguments

Determine mode and target from the command arguments:

• Default mode: standard
• Default target: conversation

Step 3: Extract Context Snapshot

Create a YAML-formatted snapshot of the current session. DO NOT include raw conversation - structure it as data:

snapshot:
  mode: [parsed mode]
  target: [parsed target]
  pal_available: [true/false from Step 1]
  pal_models: [list of models if available, empty if not]

  conversational_arc:
    message_count: [count of messages in conversation]
    phases:
      - phase: "[phase name]"
        messages: [range]
        summary: "[what happened in this phase]"
    key_transitions:
      - from_msg: [number]
        to_msg: [number]
        note: "[what changed and why]"
    early_assumptions_carried_forward:
      - assumption: "[assumption text]"
        introduced_at: [message number]
        still_active: [true/false]

  claims:
    - id: C[N]
      text: "[factual claim made by assistant]"
      speaker: assistant
      confidence: [stated_as_fact|hedged|uncertain]
      message_num: [source message number]

  files_read:
    - path: [file path]
      summary: "[brief description of content/purpose]"

  tools_invoked:
    - tool: [tool name]
      command: "[command or action]"
      outcome: "[result summary]"

  decisions:
    - decision: "[decision made]"
      rationale: "[stated reason]"

  assumptions_explicit:
    - "[explicitly stated assumption]"

Step 4: Launch Coordinator

Use the Task tool to launch a SINGLE agent:

Task: Launch red-team-coordinator agent
Agent: agents/red-team-coordinator.md
Prompt: [Include the full YAML snapshot]

Step 5: Return Output

Return the coordinator's markdown report DIRECTLY to the user.

DO NOT:

• Add commentary about the process
• Explain what the coordinator did
• Include any "I attacked X" reasoning
• Show intermediate findings

ONLY return the final sanitized markdown report.

Context Isolation Rules

This command is the FIREWALL between main session and red team work:

• Main session context stays CLEAN
• Only structured snapshot data passes to coordinator
• Only sanitized markdown report returns to user
• No adversarial prompts or attack reasoning enters main context