lo2cin4bt-teaching

name	lo2cin4bt-teaching
description	Repo-local teaching skill for lo2cin4bt. Use when explaining setup, AI usage, repo structure, frontend pages, metrics, README, Lecture Skill, AI Manual Skill, terminology, or beginner next steps from repo evidence only.
version	2.2.1
status	active
category	workflow
use_when	["Explaining lo2cin4bt setup, architecture, frontend pages, metrics, or learning material.","Updating beginner-facing Lecture or AI manuals from current repo evidence."]

lo2cin4bt Teaching Skill

Purpose

Teach the current lo2cin4bt product without reviving retired Python architecture or overstating generated evidence.

Use When

Use for onboarding, architecture explanations, Lecture updates, page walkthroughs, and terminology.

Inputs

• Learner goal and language.
• Current runtime architecture, docs, tests, configs, payloads, and frontend pages.

Workflow

Follow the Procedure below. Separate implemented behavior from general theory, cite current evidence, and end with one concrete next step.

Language Policy

1. Use response_language from the PM packet when present; otherwise infer it from the latest user message.
2. If the user writes in Chinese or asks for Chinese, write Traditional Chinese for all non-specialist wording.
3. Keep code identifiers, file paths, commands, schema keys, agent/skill names, ticker symbols, provider names, and standard finance/quant abbreviations exact. For mixed technical terms, write Chinese first with English in parentheses where useful, e.g. 夏普率 (Sharpe), 前向分析 (WFA).
4. Return response_language in the output packet. When writing repo-local reports, apply the same language policy to the report body.

Safety Notice

Teaching is educational only. It is not investment advice, trading advice, financial advice, or an instruction to trade.

Required Reads

Load only what the lesson needs:

• skills/lo2cin4bt/references/runtime-architecture.md
• README.md
• docs/ai/AI_MANUAL_SKILL.md
• docs/ai/AI_SKILL_LECTURE_GUIDE.md
• skills/lo2cin4bt/references/first-run.md
• skills/lo2cin4bt/references/frontend-pages.md
• skills/lo2cin4bt/references/metric-dictionary.md
• skills/lo2cin4bt/references/troubleshooting.md

For reader-facing Traditional Chinese Lecture copy:

• Write in plain Traditional Chinese for readers who are new to backtesting.
• Introduce technical terms as 中文（English original） on first use.
• Prefer direct verbs and concrete examples; remove optional adverbs and filler.
• Preserve config keys, commands, paths, metric names, and technical facts exactly.
• Keep this public skill self-contained. Do not depend on Company-local skills or expose local workstation paths.

Current Timeline / Rust Teaching

• Teach current strategy execution as one plain flow: strategy_run config -> Python normalization and data loading -> persistent Rust engine -> mandatory Rust result validation -> Rust metrics and PlotBundle -> outputs/app artifacts -> plotter/web on port 2424.
• Explain profiles as simpler authoring forms that all compile into this same engine. Single-asset strategies are one-asset portfolios; they are not a separate backtester.
• Explain WFA and rolling validation as explicit validation_workflow/ wrappers. They are not automatic stages for every backtest, and Parameter Matrix is candidate expansion rather than validation.
• For new examples, prefer fill_model.timing = "timeline" and explicit actions[]: offset_bars says how many bars after the signal/rebalance trigger, price is open or close, and action is enter, exit, flatten, or set_target_weights. Use 0 only for an action known before the session. Current-bar indicators, rankings, and signals must execute on a later bar through offset_bars: 1 or signal_close_for_next_bar.
• Current public examples should use timeline actions[] directly and describe one shared Rust engine. Vector precomputation and sequential portfolio accounting are internal stages of that engine, not separate public routes.
• Metrics Overview, Parameter Matrix, Backtests, and WFA read completed app-managed runs from outputs/app/latest_runs.json; they do not list every config in workspace/runs/ or workspace/wfa/. If configs exist but latest_runs.json is empty, the correct explanation is "no current completed runs yet", not "the configs cannot be read".
• If a user asks why old results disappeared after config/schema work, first check /api/app/run-center/configs, /api/app/metrics/runs, /api/app/wfa/runs, and outputs/app/latest_runs.json.

Procedure

1. Identify the learner goal.
2. Name the repo evidence used.
3. Explain implemented lo2cin4bt behavior separately from general trading theory.
4. Define technical terms at first use.
5. Mention what a page or artifact can prove and cannot prove.
6. For complete beginners, provide a learning goal, prerequisites, expected result, common problems, and a copyable checklist.
7. When presenting hands-on work, separate the manual route from the Agent/Skill route. Explain ProjectManager, WorkAgent, and Skill before using those names.
8. End with one concrete next file, page, command, or artifact.
9. For Lecture copy, use documentation mode, keep technical facts and names, and remove optional adverbs. Record any meaning-critical exception; the expected exception list is empty.

Repo-Local Report Output

When ProjectManager asks for a durable repo-local report, write Markdown under workspace/reports/agents/<agent_name>/YYYY-MM-DD_<short-topic>.md and return repo_agent_report_path in the output packet. Use this for WorkAgent and private snapshot reports; keep Company-level closeout separate unless ProjectManager explicitly asks for it.

Stop Conditions

Stop when the request needs trading advice, unsupported repo behavior, unavailable private credentials, paid data access, or old artifacts that conflict with current contracts.

Output Format

response_language:
lesson_goal:
repo_evidence:
plain_explanation:
terms:
not_generated_or_not_applicable:
next_step:
repo_agent_report_path:
not_trading_advice_notice:

Outputs

• Beginner-readable, evidence-backed lesson or updated teaching artifact.
• Clear not generated, not applicable, and safety boundaries.

Path Rules

• Lecture source: <repo-root>/Lecture/.
• Product docs and skills remain under the Repo root.
• Teaching reports belong under the assigned repo-local and Company report roots.

Validation

uv run --locked --exact --group dev python -m pytest tests/test_agent_skill_lecture_alignment.py tests/test_lecture_contracts.py -q

Pass criteria: local links resolve, UTF-8 is valid, lessons teach the shared Rust route and explicit validation workflows, and the beginner learning map, glossary, two operation routes, checklist, copy controls, and progress components remain present. Every changed Lecture HTML page must also follow the plain-language rules above without optional adverbs or internal environment references.

Fail action: correct the teaching source; never document a compatibility exception for retired behavior.

Report Requirements

Record lessons changed, source evidence, terminology decisions, validation result, and any feature that remains optional or not implemented.