Instructions file CodexOpenCode
AGENTS.md instructions for LightDevCoder/skills, covering first-party skills repository maintenance contract, ownership boundary, repository shape and sources of truth, authoritative references and maintenance rules and invocation direction.
Agent
How the engineering skills should consume this repo's domain documentation when exploring the codebase.
Agent
Issues and specs for this repo live as markdown files in .scratch/.
Agent
The skills speak in terms of five canonical triage roles. This file maps those roles to the actual label strings used in this repo's issue tracker.
Skill Claude CodeCodex
Map the current Agent Host's evidenced models and execution capabilities to a safe, role-clear execution plan. Use before complex multi-agent work when routing, reviewer independence, ownership, or host limits affect the result.
Skill Claude CodeCodex
Act as the Light workflow advisor, navigator, and router. Understand what the user is trying to do, inspect the actual project/host evidence and Skill catalog, reason about the best next Light Skill, explain why, wait for user approval, then honor the accepted Skill's invocation policy with host-aware transition. Also…
Skill Claude CodeCodex
Run a lightweight standalone clarification session for an ambiguous idea, requirement, plan, or process. Use only when the user explicitly invokes $clarify to start; normal replies then continue until shared understanding is confirmed, without generating a formal specification or auto-chaining another workflow.
Skill Claude CodeCodex
Review the diff since a fixed point along Standards and Spec axes in parallel sub-agents, reporting findings without editing the target or deciding the final verdict. Use when the user wants to review a branch, PR, or work-in-progress diff, or when review-loop or project-review requests a specialist software check.
Skill Claude CodeCodex
Plan a large, multi-session, decision-heavy effort as a persistent decision map of tickets. Use only when the user explicitly invokes $decision-map; it maintains .scratch/ /map.md and child decision tickets until fog clears, then hands off to project-spec.
Skill Claude CodeCodex
Diagnosis loop for hard bugs and performance regressions. Use when the user says "diagnose"/"debug this", or reports something broken/throwing/failing/slow.
Skill Claude CodeCodex
Explain any topic, code, concept, or error tailored to a specific audience's level of understanding. Use this skill whenever the user says 'explain like I am', 'ELI5', 'explain this to my', 'break this down for', 'dumb it down', 'simplify this for', or asks you to explain something to a specific person or audience…
Skill Claude CodeCodex
Review a bounded non-specialist artifact against supplied requirements and return normalized, read-only findings. Use when no accepted specialist reviewer is more appropriate.
Skill Claude CodeCodex
Compact the current conversation into a handoff document for another agent to pick up.
Skill Claude CodeCodex
Rewrite AI-sounding text so it reads naturally without changing what it says. Use when editing or reviewing prose for inflated claims, sales language, vague sources, repetitive structure, stock AI words, passive voice, filler, or chatbot artifacts. Works for English and Chinese text (去除中英文文本中的 AI 写作痕迹，使行文自然、更像人写).…
Skill Claude CodeCodex
Execute one bounded, already-decided work item — code, document, configuration, research artifact, Skill, or generic project task — by inspecting relevant context, offering agent-config routing when materially useful, verifying locally, and handing the result to review-loop with the appropriate specialist reviewer.
Skill Claude CodeCodex
Pick up and execute one task from a Light-Kanban board, then return it for human confirmation. Use when a scheduled agent wakes and must process queued Light-Kanban work — resume owned in-progress tasks and Request Changes feedback first, claim new To Do work with a stable agent identity, validate the workspace…
Skill Claude CodeCodex
Design and initialize a maintainable knowledge base through a knowledge-base-specific interview. Use only when the user explicitly invokes kb-init to create, rebuild, structure, or initialize a knowledge base, wiki, reference library, research archive, operational knowledge system, or similar long-term knowledge…
Skill Claude CodeCodex
Language-learning tutor for any target language. Creates 30-minute daily lessons on grammar, speaking, listening, vocabulary, pronunciation, or writing; turns word lists into flashcards with usage examples and memory tips; holds natural conversation practice as a native speaker with gentle corrections; decodes grammar…
Skill Claude CodeCodex
Turn conversations, transcripts, project notes, folder workflows, documentation, or other sources into reusable agent skills and Skill Creator compatible instructions. Use when a user asks to create, update, distill, or "learn" a workflow as a skill; preserve repeated operating methods for future AI agents; encode…
Skill Claude CodeCodex
Route and govern manuscript engineering from small notes and structured reports to large manuals, books, multilingual editions, and reproducible multi-format deliverables. Use when an agent must assess document scope and risk, create or resume a writing project, manage sources and semantic batches, coordinate explicit…
Skill Claude CodeCodex
Clarify a real project's unresolved decisions from inspected project facts and return a bounded handoff artifact. Use only when the user explicitly invokes $project-clarify after readiness or when a project needs formal clarification; it does not create a SPEC, tickets, or automatically start another user-invoked…
Skill Claude CodeCodex
Bootstrap the stable Light project and tracker contracts consumed by downstream Project Skills, preserving manual content and supporting safe reruns. Use only when a user explicitly invokes $project-init for a new or incompletely initialized repository.
Skill Claude CodeCodex
Determine whether a completed project satisfies its approved target by freezing an acceptance baseline, composing reviewers, and issuing a final PASS/FAIL/BLOCKED. Use when project delivery needs final acceptance.
Skill Claude CodeCodex
Turn already-clarified outputs from project-clarify or decision-map into a formal project SPEC without reopening an interview. Use only when the user explicitly invokes $project-spec; it returns to project-clarify when a blocking user decision remains and hands off to project-tickets when ready.