Aicrew Dev

aicrew /dev (Codex)

Codex does not support slash commands. Use this skill when the user asks for "/dev" or a full SDLC pipeline.

Token foundation (mandatory)

/dev, /fix, and /quick all share the same 11-capability token-saving stack — only pipeline depth differs. Full reference: ~/Agents/docs/token-foundation.md. Stack: graph-first (codebase-memory-mcp), speculative Scout → verify (SCOUT schema, two-model routing), Karpathy guardrails, layered guardrails (guardrails-taxonomy.md), context-economy read policy, security-guard.py hooks, .ai/state checkpoints, /compact between phases, /handoff on tool switch, optional context-mode + token-optimizer-mcp, caveman default output. For /dev, Scout opens Phase 1 Research before any Glob/Grep/Read; re-scout between phases if context grew.

Default output

Caveman/lean style by default. See ~/Agents/agents/caveman.md and ~/Agents/agents/context-economy.md. /normal or /lean off restores verbose.

Source of truth:

• ~/Agents/commands/dev.md
• Project overrides in .ai/skills/ and repo AGENTS.md (if present)

Token foundation (mandatory — all phases):

• Graph-first research: codebase-memory-mcp (searchgraph → tracepath → getcodesnippet) before any Grep/Read
• Speculative context: Scout pass (cheap model) at start of Phase 1; emit SCOUT: schema; verify before main trace
• Layered guardrails: security-guard.py (input) → scope lock (Phase 0) → karpathy-guardrails (Phase 4) → security-reviewer (Phase 6)
• Context economy: always on; /compact between phases; /lean amplifies
• Two-model routing: Scout on haiku/mini; Research+Implement on sonnet/opus
• See: ~/Agents/docs/token-foundation.md

Workflow summary:

1. Intake: clarify bug/feature/refactor, acceptance criteria, scope, risks, test plan.
2. Research: Scout pass (graph-first) → verify SCOUT schema → confirm root cause or key code paths.
3. Brainstorm: 3 alternatives with trade-offs (features/refactors).
4. Design: interface spec, contract checks, over/under-engineering flags.
5. Implement: load karpathy-guardrails; TDD first (RED -> GREEN -> REFACTOR per acceptance criterion).
6. Tests: targeted automated tests + smoke path.
7. Security: changed files only, no false positives.
8. Audit: if project defines domain audit checks.
9. Conclude: summary, tests run, risks, commit message.

Checkpoints: pause for user confirmation at intake, design, and before concluding.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: AKSoftCode
• Source: AKSoftCode/aicrew
• License: MIT
• Homepage: https://github.com/AKSoftCode/aicrew

Install and usage instructions live in the source repository linked above.