Parity Matrix — SubQ Code vs Claude Code vs oh-my-pi

Head-to-Head Matrix

Has Shipped Partial Primitives exist Planned On roadmap None Not available

Capability

SubQ Code

Claude Code

oh-my-pi

Tier 1 — Architectural Foundations

R1: Prompt Cache Boundary

Partial

Has

None

R2: Lifecycle Hooks

Partial

Has

R3: Subagent Spawning

Partial

Has

Tier 2 — Agent Intelligence

R4: Risk-Based Authorization

Planned

Has

None

R5: Persistent Memory

Planned

Has

R6: Verification Agent

Partial

Has

None

R7: Model-Specific Variants

Planned

Has

None

Tier 3 — Operational Excellence

R8: Numeric Length Anchors

Partial

Has

None

R9: MCP Integration

Planned

Has

R10: Token Budget Mode

Planned

Has

None

R11: Autonomous Mode

Partial

Has

R12: Feature Flags

Partial

Has

Tier 4 — Beyond Parity (oh-my-pi Innovations)

R13: TTSR (Streamed Rules)

Planned

None

Has

R14: Autonomous Memory Pipeline

Planned

None

Has

R15: Session Tree

Partial

None

Has

R16: Model Roles

Partial

None

Has

R17: Cross-Agent Rule Discovery

Partial

None

Has

R18: Extension API

Partial

Has

R19: Task Agent Swarm

Partial

Has

R20: Intelligent Commit Tool

Planned

None

Has

SubQ Advantages (Already Leading)

Six capabilities no competitor matches. None are modified by the parity plan—these are preserved advantages.

FFF Search Cursor pagination, multi_grep, blast_radius analysis. Vastly superior to basic Glob/Grep.

Context Packing LLM-driven context_pack with query synonyms. Smarter than brute-force Explore agent.

Behavioral Nudges System reminders engine: bash spiral, plan drift, 5 reactive events.

Quality Measurement Leverage analysis, prompt ROI, orchestration scoring. Data-driven feedback.

Plan-Driven Workflow Plannotator with interview depths, TUI review. Structured over reactive.

Cross-Agent Intelligence 12-agent parser ecosystem. Multi-agent parsing no one else has.

Claude Code Advantages (Gaining)

12 capabilities Claude Code ships today. After this roadmap, SubQ Code has all 12.

R1

Prompt Cache Boundary

Static/dynamic split with cache_control markers. ~80% hit rates.

Phase 1

R2

Lifecycle Hooks

18 hook events with shell command execution at lifecycle points.

Phase 2

R3

Subagent Support

Fork + typed variants, background task notification.

Phase 3

R4

Risk-Based Reasoning

3-category taxonomy in system prompt. Self-assessment per action.

Phase 4

R5

Persistent Memory

Manual MEMORY.md, file-based storage, auto-recall on start.

Phase 5

R6

Verification Agent

Adversarial independent checking of non-trivial work.

Phase 6a

R7

Model-Specific Variants

Per-family variants with feature gating, comment suppression.

Phase 4

R8

Output Control

Numeric length anchors proven to reduce output tokens.

Phase 1

R9

MCP Integration

Full protocol with delta optimization for late connects.

Phase 6b

R10

Token Budget Mode

Hard minimum target, auto-continue on early stop.

Phase 6a

R11

Autonomous Mode

Sleep tool, terminal focus awareness, cost guardrails.

Phase 6b

R12

Feature Flags

Build-time DCE, runtime evaluation, remote resolution.

Phase 1Already built

oh-my-pi Innovations (Surpassing)

8 capabilities pioneered by oh-my-pi that neither Claude Code nor SubQ Code has today. After this roadmap, SubQ has all 8.

R13

TTSR — Streamed Rules

Zero-context-cost rules via stream monitoring with regex triggers. Abort/retry on match.

NovelPhase 2

R14

Autonomous Memory Pipeline

LLM-powered extraction + consolidation. Auto-generated skills. SQLite job queue.

Supersedes R5Phase 5

R15

Session Tree

Fork, resume, cross-project. Sessions as tree, not linear. HTML export.

NovelPhase 5

R16

Model Roles

Default, smol, slow, plan, commit. Auto-discovery. 30% cost reduction.

NovelPhase 3

R17

Cross-Agent Rule Discovery

Discover from 7 agent config formats. Normalize, deduplicate, priority-based.

NovelPhase 2

R18

Extension API

30+ events, tool/command/renderer registration, inter-extension messaging.

Supersedes R2Phase 2

R19

Task Agent Swarm

100 parallel isolated tasks. Git worktree/fuse-overlay/ProjFS. Merge strategies.

Supersedes R3Phase 3

R20

Intelligent Commit Tool

Agentic git inspection, split commits, hunk-level staging, conventional format.

NovelPhase 6b

Verdict

After this roadmap: SubQ Code will have all 12 Claude Code advantages + all 8 oh-my-pi innovations + 6 unique SubQ capabilities = 26 capabilities no other single coding agent offers. Claude Code maxes out at 14. oh-my-pi at 14. SubQ Code ships with all of them—plus 6 that neither competitor has.

Source

SubQ Code

Claude Code

oh-my-pi

Claude Code parity (R1–R12)

oh-my-pi innovations (R13–R20)

Unique SubQ capabilities

Total