Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add youseiushida/benkyo --skill benkyo-treatment-shiftgit clone --depth 1 https://github.com/youseiushida/benkyoWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/youseiushida/benkyo/benkyo-treatment-shift)<a href="https://agentmods.dev/skills/youseiushida/benkyo/benkyo-treatment-shift"><img src="https://agentmods.dev/badge/skills/youseiushida/benkyo/benkyo-treatment-shift/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/youseiushida/benkyo/benkyo-treatment-shift"><img src="https://agentmods.dev/badge/skills/youseiushida/benkyo/benkyo-treatment-shift.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00000 | $0.03022 |
| Opus 5 | $0.00000 | $0.01511 |
| Sonnet 5 | $0.00000 | $0.00604 |
| Haiku 4.5 | $0.00000 | $0.00302 |
Grade A, and why
benkyo-treatment-shift scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 12d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 216 lines — stays where its author put it; the contents beside it link to each section on GitHub.
benkyo-treatment-shift: changing depth of engagement
This skill handles the operation of changing a concept's treatment within a project — moving it from "use as a tool" (blackbox) to "really understand" (whitebox), or vice versa. Both directions have specific protocols.
The decision when to shift is part of this skill. The actual ongoing teaching is in benkyo-tutoring.
Scope of this skill
This skill decides and performs the treatment shift — confirming intent, updating the graph, setting the treatment, and logging the change. It does NOT conduct the ensuing explanation or problem-solving. After any shift, defer immediately to benkyo-tutoring to continue the session.
Cardinal vocabulary rule
Internal terms (blackbox, whitebox, treatment, cut, commit, release, prereq, node, edge, graph, event, log, record, schema, JSON, metadata, treatment_changed, toggle, split, granularity) are NEVER spoken to the learner. Translate:
- commit / 「上に上げる」 → 「ここを掘り下げる」「ちゃんと理解する」
- release / 「下に降ろす」 → 「公式で済ます」「ツールとして使う」「飛ばす」
- blackbox → 「公式扱い」「使えれば OK」
- whitebox → 「理解する対象」
When you internally execute benkyo treatment set ..., what you say out loud is "じゃあここは公式覚えて使う形に切り替えるね" (or similar), NEVER "treatment を blackbox にしました".
Internal IDs are also forbidden: never write c1, p1, prj1 etc. in learner-facing text. Refer to concepts by their natural-language name or content.
See ../_benkyo-shared/references/nl-to-cli.md for the full translation table.
Trigger types
Direct signals (from learner)
| Learner says | Direction |
|---|---|
| 「理解したい」「腹落ちしたい」「導出やりたい」「なぜそうなる？」「もっと深く」 | commit (blackbox → whitebox) |
| 「公式覚えれば」「暗記でいい」「ざっくり」「飛ばして」「もういい」 | release (whitebox → blackbox) |
When a direct signal is clear, confirm intent briefly then execute.
Indirect signals (inferred by tutor)
For commit:
- Repeated "なんで？" or meaning-questions about the same concept (≥ 2 times in a session)
- Misapplying the blackbox reference (e.g., wrong formula choice, wrong substitution pattern)
- Failed transfer: solves the original problem but fails when conditions change slightly
- "Which formula here?" repeated — suggests not understanding applicability conditions
What ships with it
2 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 12d ago First seen · 216 lines · 0 tokens per session scan A c17df709dad8
benkyo-treatment-shift is a skill published in the GitHub repository youseiushida/benkyo (2 stars, last pushed 4mo ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 3,022 tokens. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
automating-devops
DevOps knowledge reference covering Git workflows, testing strategies, DevSecOps, release pipeline orchestration (release.yml, multi-arch images, cosign integration), CI/CD pipelines, database management, observability, and performance optimization. Use when working with Git, CI/CD, release pipelines, ghcr image…
cap-commit-push
Use when the user sends cap or asks for a fast commit-and-push shortcut for current repository changes.
versioning-by-impact
Use when completing a task and deciding whether repository changes require a semantic version bump and changelog entry before commit/push.
codex-hud
Show Codex workspace context with local config, git state, hooks, and project hints. Standalone mode can show an expanded snapshot; patched Codex TUI footer mode is single-line. Use when the user says "codex-hud", "show Codex HUD", "status HUD", "workspace snapshot", or asks for richer Codex context.
agent-rollback
Use the agent-rollback MCP server or ar CLI to snapshot the workspace, list/show/diff checkpoints, and roll back unwanted changes. Use this skill whenever the user mentions checkpoints, rollback, undo, revert, restore, "go back to", "what changed", "I broke something", "this isn't working", or asks to inspect or…
shipping-changes
End-to-end change-shipping closed loop for non-trivial work — research → proposal doc → phased guarded implementation → PR self-review → harden → merge. Use when taking a substantial change from idea to merged PR, when a task needs a written proposal before code, when implementing in reversible phases behind a guarded…