AKSoftCode aicrew /fix (Codex)
Use for fast bug fixes. This mirrors the /fix command but runs inline in Codex.
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 /fix, Scout opens Phase 1 Bug Analysis before the bug-analyst deep dive.
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/fix.md- Project overrides in
.ai/skills/and repoAGENTS.md(if present)
Token foundation (mandatory — all phases):
- Graph-first: codebase-memory-mcp (search_graph → trace_path → get_code_snippet) before any file read
- Speculative context: Scout pass at start of Phase 1 Bug Analysis; emit SCOUT: schema; verify before bug-analyst deep dive
- Layered guardrails: security-guard.py (input) → karpathy-guardrails (Phase 2 implement) → security-reviewer (Phase 4)
- Context economy: always on; slice reads only during Scout
- Two-model routing: Scout on haiku/mini; fix on sonnet
- See: ~/Agents/docs/token-foundation.md
Workflow summary:
- Ask the 3 clarifying questions (symptom, expected behavior, repro).
- Scout pass (graph-first) → emit SCOUT: schema → verify → bug-analyst deep trace → confirm root cause.
- Load karpathy-guardrails; write the smallest failing test or reproducible check.
- Implement the minimal fix to make it pass.
- Run targeted tests + smoke path.
- Security review on changed files.
- Conclude with summary, tests run, and any risks.