System Prompt (Complete Reconstruction)

Full document | Source: src/constants/prompts.ts, src/utils/systemPrompt.ts

5-Layer Priority System

Priority 0: Override (loop mode, testing)        ← highest
Priority 1: Coordinator (multi-worker orchestration)
Priority 2: Agent (subagent definitions)
Priority 3: Custom (--system-prompt flag)
Priority 4: Default (standard Claude Code)       ← lowest
         + appendSystemPrompt (always appended unless override)

What the Default Prompt Actually Says

The system prompt is 9 major sections:

1. Identity — "You are an interactive agent that helps users with software engineering tasks"
2. System Rules — Tool execution, permission modes, prompt injection detection
3. Task Execution — The most detailed section:
   • Read code before modifying (mandatory)
   • Don't add unrequested features/refactoring
   • Don't create helpers for one-time operations
   • "Three similar lines of code is better than a premature abstraction"
   • Avoid OWASP Top 10 vulnerabilities
4. Careful Actions — Destructive operations need user confirmation (rm -rf, force push, etc.)
5. Tool Usage — Use dedicated tools (Read/Edit/Glob/Grep) instead of Bash
6. Tone — No emojis, concise, file_path:line_number format
7. Output Efficiency — "If you can say it in one sentence, don't use three"
8. Cache Boundary — __SYSTEM_PROMPT_DYNAMIC_BOUNDARY__ splits static/dynamic parts
9. Environment — CWD, platform, model name, knowledge cutoff

Cache Optimization

The prompt is split by __SYSTEM_PROMPT_DYNAMIC_BOUNDARY__:

• Before boundary (static): Cached globally, saves tokens on repeat calls
• After boundary (dynamic): User/session-specific, recalculated each time (memory, MCP, language, etc.)

Model Knowledge Cutoffs

| Model | Display Name | Cutoff | |-------|-------------|--------| | claude-opus-4-6 | Claude Opus 4.6 | May 2025 | | claude-sonnet-4-6 | Claude Sonnet 4.6 | August 2025 | | claude-haiku-4-5 | Claude Haiku 4.5 | February 2025 |

87 Hidden Feature Flags

Full document | Source: src/commands.ts, various

These are compile-time switches via bun:bundle feature(). Most are dead-code-eliminated in the public npm build.

Core Codenames

| Codename | What It Does | Cross-file References | |----------|-------------|----------------------| | KAIROS | Autonomous assistant platform (assistant mode, brief, channels, cron, webhooks) | 210 files | | PROACTIVE | Proactive task planning and automation | Linked to KAIROS | | COORDINATOR_MODE | Multi-agent orchestration with workers | 45 files | | VOICE_MODE | Voice input/output with cross-platform binaries | 38 files | | BUDDY | AI companion sprite with animation system | 14 files, 1,298 lines | | ULTRATHINK | Extended deep reasoning mode | Compile-time | | ULTRAPLAN | Super planner with selection dialog | Compile-time | | TORCH | Inference enhancement | Compile-time | | BRIDGE_MODE | Mobile/web remote control | Compile-time | | CHICAGO_MCP | Computer-use MCP server | Internal |

Context & Compression

| Flag | Purpose | |------|---------| | CONTEXT_COLLAPSE | Smart context folding | | REACTIVE_COMPACT | Reactive session compression | | CACHED_MICROCOMPACT | Micro cache compression | | HISTORY_SNIP | History message fragmentation | | BG_SESSIONS | Background session management | | AWAY_SUMMARY | Auto-summary when user is away |

Obfuscated Runtime Config (tengu_* gates)

Runtime flags use tengu_ + random word pairs to deliberately hide their purpose:

tengu_frond_boric       → Analytics killswitch (Datadog/FirstParty)
tengu_passport_quail    → Memory extraction gate
tengu_moth_copse        → Memory extraction enable
tengu_bramble_lintel    → Memory extraction frequency
tengu_cicada_nap_ms     → Background refresh throttle
tengu_slate_prism       → Connector text summarization
tengu_amber_json_tools  → JSON tool format (token-efficient)
tengu_tool_pear         → Structured output (strict tools)

15 Hidden Slash Commands

Source: src/commands.ts

These require specific feature flags to appear:

| Command | Function | Required Flag | |---------|----------|---------------| | /assistant | Assistant mode | KAIROS | | /brief | Brief messages | KAIROS_BRIEF | | /bridge | Remote control bridge | BRIDGE_MODE | | /voice | Voice mode | VOICE_MODE | | /buddy | AI companion sprite | BUDDY | | /ultraplan | Super planner | ULTRAPLAN | | /torch | Inference boost | TORCH | | /workflows | Workflow scripts | WORKFLOW_SCRIPTS | | /fork | Subagent branching | FORK_SUBAGENT | | /peers | Peer messaging | UDS_INBOX | | /proactive | Proactive planning | PROACTIVE | | /force-snip | Force history snip | HISTORY_SNIP | | /subscribe-pr | PR subscriptions | KAIROS_GITHUB_WEBHOOKS | | /remote-setup | Remote CCR setup | CCR_REMOTE_SETUP | | /remote-control-server | Remote control server | DAEMON + BRIDGE_MODE |

25 Internal-Only Commands

Condition: USER_TYPE === 'ant' && !IS_DEMO

| Command | Function | |---------|----------| | /bughunter | Automated bug detection | | /good-claude | Model evaluation tool | | /commit | Hidden git commit | | /commit-push-pr | One-click commit → push → create PR | | /ctx-viz | Context visualization | | /break-cache | Force cache invalidation | | /mock-limits | Simulate rate limits | | /reset-limits | Reset rate limits | | /ant-trace | Anthropic internal tracing | | /perf-issue | Performance diagnostics | | /debug-tool-call | Tool call debugger | | /agents-platform | Agents platform management | | /autofix-pr | Auto-fix PR | | /backfill-sessions | Backfill session data | | /share | Session sharing | | /summary | Session summary | | /onboarding | Onboarding flow | | /init-verifiers | Initialize verifiers | | /bridge-kick | Kick bridge connection | | /version | Version display | | /oauth-refresh | OAuth token refresh | | /env | Environment variables | | /issue | Issue tracking | | /teleport | Session teleport | | /tags | Session tags |

Undercover Mode

Full document | Source: src/utils/undercover.ts

When Anthropic employees (USER_TYPE=ant) work in public repositories, Claude Code automatically hides all internal information.

How It Works

isUndercover() check:
  USER_TYPE !== 'ant' → always OFF (external users unaffected)
  USER_TYPE === 'ant' →
    CLAUDE_CODE_UNDERCOVER=1 → forced ON
    Repo in internal allowlist → OFF
    Repo is public/unknown → ON (safe default)

There is NO force-OFF option. If Claude Code can't confirm it's an internal repo, undercover stays ON.

What Gets Injected

## UNDERCOVER MODE — CRITICAL

You are operating UNDERCOVER in a PUBLIC/OPEN-SOURCE repository.
NEVER include in commit messages or PR descriptions:
- Internal model codenames (Capybara, Tengu, etc.)
- Unreleased model version numbers
- Internal repo or project names
- The phrase "Claude Code" or any mention that you are an AI
- Co-Authored-By lines or any other attribution

Internal Repo Allowlist (22 private repos)

anthropics/claude-cli-internal, anthropics/anthropic, anthropics/apps,
anthropics/casino, anthropics/dbt, anthropics/dotfiles,
anthropics/terraform-config, anthropics/hex-export, anthropics/feedback-v2,
anthropics/labs, anthropics/argo-rollouts, anthropics/starling-configs,
anthropics/ts-tools, anthropics/ts-capsules, anthropics/feldspar-testing,
anthropics/trellis, anthropics/claude-for-hiring, anthropics/forge-web,
anthropics/infra-manifests, anthropics/mycro_manifests,
anthropics/mycro_configs, anthropics/mobile-apps

Source: src/utils/commitAttribution.ts lines 30-75

Cost Optimization

Full document | Sources: src/utils/modelCost.ts, src/utils/context.ts, src/services/compact/

Model Pricing (per million tokens)

| Model | Input | Output | Cache Read | Cache Write | |-------|-------|--------|------------|------------| | Haiku 4.5 | $1 | $5 | $0.10 | $1.25 | | Sonnet 4.6 | $3 | $15 | $0.30 | $3.75 | | Opus 4.6 | $5 | $25 | $0.50 | $6.25 | | Opus 4.6 Fast | $30 | $150 | $3.00 | $37.50 |

Fast mode is 6x the price. Use it only for short, time-sensitive tasks.

Source: src/utils/modelCost.ts lines 36-69

10 Money-Saving Tips

| # | Tip | Source Evide