Debugging Modularization (SoT, agent-neutral)

Source: dev-pabcd/references/debugging-modularization.md

Debugging Modularization (SoT, agent-neutral)

> Source-of-truth contract for runtime-specific debugging references shared across

> the pabcd_initiative harnesses (codexclaw, cli-jaw, jawcode). This file names the

> structure and content contract; each harness ships the references in its own

> skill directory.

Status

Harness	Location	Status
codexclaw	`plugins/codexclaw/skills/dev-debugging/references/runtimes/` + `tools/`	SHIPPED (2026-07-07, 9 runtimes + js/ framework tier + 1 tool)
cli-jaw	`skills_ref/dev-debugging/references/runtimes/` + `tools/`	SHIPPED (2026-07-07, 9 runtimes + js/ framework tier + 1 tool, synced)
jawcode	n/a (fork upstream)	NOT STARTED (9 runtimes available for port) — port when jawcode's dev-debugging gains references

Structure Contract

references/
  runtimes/
    node.md     # Node.js / tsx / Bun / Deno
    js/         # JS framework tier (node.md is the hub)
      nextjs-react.md   # Next.js 16 / React 19 / RSC
      vite-vitest.md    # Vite 8 / Vitest 4 / HMR / build failures
      node-backend.md   # Express 5 / Fastify 5 / NestJS 11
    python.md   # CPython 3.9+
    rust.md     # cargo / tokio
    go.md       # Delve / pprof
    c-cpp.md    # sanitizers / GDB / LLDB / Valgrind
    jvm.md      # jcmd / JFR / JDWP / Kotlin / GraalVM
    swift.md    # LLDB / Instruments / simulator CLI
    ruby.md     # debug gem / rdbg / Rails
    beam.md     # IEx / Observer / :dbg / recon
  tools/
    playwright.md   # browser-surface deterministic repro

Each runtime file MUST include: Phase 0 environment detection, attach/launch

recipes, the runtime's debugging hierarchy (lightest tool first), a

silent-failure pattern table, and a cleanup section. Lineage: lazycodex

debugging/references/runtimes/ + gpt-5.5 Tier-2 external research round.

Slop Cleanup Checklist (REVIEW-SLOP-01)

Shipped in dev-code-reviewer SKILL.md on both codexclaw and cli-jaw. 9

categories adapted from lazycodex remove-ai-slops: stylistic (obvious

comments, over-defensive, excessive complexity), structural (needless

abstraction, boundary violations, oversized >250L as smell not mandate),

hidden cost (perf equivalences, scope leaks), coverage (missing behavior

tests). Safety invariant: green tests before removing code.

Sources

lazycodex debugging/references/runtimes/{node,python,rust,go}.md
lazycodex remove-ai-slops/SKILL.md (9-category taxonomy)
gpt-5.5 research round (2026-07-07): Node 22+ inspect-wait, Python 3.14

PEP 768, Vitest 4 debugging, Miri/tokio-console, Delve/pprof, Playwright

agent integration