app-design-thinking
Design the app mechanism and build pipeline for the produced app — the app-phase analog of [[schema-design]]. Use this skill whenever the knowledge phases are…
Execute John's high-volume vertical phases with native Codex subagents over the durable `.john/runs` and event contracts. Use for per-chunk extraction, coverage, grounding, per-entry generation, large uniform fan-out, retries, reconciliation, cancellation, or whenever shared
$ npx -y skills add kitchen-engineer42/joharnessburg --skill codex-vertical-workflows --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/codex-vertical-workflowsContext preview
The summary Claude sees to decide when to auto-load this skill.
Execute John's high-volume vertical phases with native Codex subagents over the durable `.john/runs` and event contracts. Use for per-chunk extraction, coverage, grounding, per-entry generation, large uniform fan-out, retries, reconciliation, cancellation, or whenever shared
name: codex-vertical-workflows description: Execute John's high-volume vertical phases with native Codex subagents over the durable `.john/runs` and event contracts. Use for per-chunk extraction, coverage, grounding, per-entry generation, large uniform fan-out, retries, reconciliation, cancellation, or whenever shared John guidance mentions Claude dynamic workflows but the active provider is Codex.
Keep the phase contract provider-neutral: immutable indexed inputs, stable item IDs, typed receipts, unique events, audit barriers, and a verified checkpoint. Use `john_run.py` to create and reconcile the durable ledger before treating agent returns as completion.
1. Create the run from a deterministic work CSV. 2. Dispatch ordinary native subagents in bounded waves; default concurrency is 6 and workers remain leaves (`max_depth=1`). 3. Use experimental `spawn_agents_on_csv` only when capability detection says it is available and the rows are uniform. Both engines consume the same manifest and receipt contract. 4. Record every attempt receipt, then reconcile. A successful thread is not a completed item until its referenced events exist, parse, match run/item identity, and contain the required terminal event. 5. Run extraction verification, coverage, grounding, adjudication, and typed reduction as separate stages with barriers. For extraction, invoke `reduce_events.py --require-extraction-audits` before phase advancement.
Use `john_run.py status`, `retry-csv`, and `cancel` for recovery. Never assign deterministic indexing to one agent, and never let parallel code workers edit the same files without disjoint ownership or an isolated worktree.
中文版: README_ZH.md John turns unstructured source material into a working knowledge-dense app. It keeps knowledge engineering and app building in one durable run, coordinates large per-entry fan-outs, and leaves auditable events and checkpoints on disk.
Design the app mechanism and build pipeline for the produced app — the app-phase analog of [[schema-design]]. Use this skill whenever the knowledge phases are…
Bundle a finished John workspace from Codex. Use when the user wants to archive, package, hand off, or preserve a John project, or wants the Claude command…
Break parsed markdown into a tree of progressively-disclosed chunks for downstream extraction. Use this skill whenever a phase needs to work on per-chunk…
Apply deterministic quality checks to the code John produces — catch the 80% of issues (leaked API keys, hardcoded prod URLs, broken imports, missing…
Generate John's process scorecard, auditor manifests, and shareable run report from a Codex project using John's provider-neutral scripts. Use when the user…
Activate a Hamster-built or otherwise applied John template for Codex in the current project. Use when a merged template plugin already exists, when the user…