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 agentmods add skills/vthinkdeveloper/vthink-agent-toolkit/work-towards-alignmentnpx skills add vthinkdeveloper/vthink-agent-toolkit --skill work-towards-alignmentgit clone --depth 1 https://github.com/vthinkdeveloper/vthink-agent-toolkitWhat 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 | $0.00127 | $0.04641 |
| Opus 5 | $0.00063 | $0.02320 |
| Sonnet 5 | $0.00025 | $0.00928 |
| Haiku 4.5 | $0.00013 | $0.00464 |
Grade A, and why
work-towards-alignment scanned grade A with 1 finding 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 2d 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.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
- For "we accept the `*/*` known limitation": *what `*/*` means in HTTP `Accept` headers, default browser/curl/HttpClient behavior, RFC 7231 media-type matching* How it starts
The opening of the file, as written. The whole thing — 296 lines — stays where its author put it; the contents beside it link to each section on GitHub.
work-towards-alignment
Run a collaborative alignment session between you and the user about a plan, design, or PR review. The session ends only when the user explicitly signals alignment — not when you decide.
This skill is built on the assumption that alignment requires genuine understanding, not nodding-through. Each load-bearing claim depends on foundational knowledge the user may or may not already have. The skill runs in two phases:
- Phase A (chat-only): probe the user's foundational knowledge in chat, teach where needed, and record what we converged on in working memory.
- Phase B (write artefacts + walk claims): create the alignment doc reflecting what Phase A established, then present each claim with the foundation already in place.
This ordering matters: the alignment doc should reflect the conversation, not predict it.
Core principles
- Collaborative, not autonomous. No Stop hooks. No background loops. No pestering.
- User-paced. They drive. Go silent between turns. Respond when prompted.
- Teach before aligning. Each item has prerequisite knowledge the user must hold to evaluate it. Probe first in chat; teach if needed; only then write the doc and present the claim.
- Doc reflects, not predicts. Phase A happens before any file is created. The alignment doc gets written once the basics conversation has surfaced what the user knows. Prerequisites the user already understood appear as
(known); prerequisites we taught appear with a summary in the Q&A log. - Durable. Once written, all progress lives in
docs/<topic>-alignment.html(the primary, living artefact). The session survives across Claude Code sessions because the HTML is the canonical source of state. - User has final say. They can waive items, override, skip Phase A entirely ("just write the doc"), or end the session early. Never declare alignment unilaterally.
- HTML is the primary artefact. Per Anthropic's "The unreasonable effectiveness of HTML": humans don't read 100-line markdown files, but they DO scan rich HTML with tables, color-coded badges, callouts, and interactive elements. Alignment sessions accumulate exactly that kind of structured content (checklist + Q&A + decisions + basics badges) — HTML is the right medium, markdown is the wrong one. A frozen
.mdsnapshot may exist as a git-diffable audit trail but is NEVER the live working document.
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.
- 2d ago First seen · 296 lines · 127 tokens per session scan A 8775ab5a2dcb
work-towards-alignment is a skill published in the GitHub repository vthinkdeveloper/vthink-agent-toolkit (2 stars, last pushed 1mo ago), licensed MIT. It adds 127 tokens to every session and 4,641 once invoked, about $0.0006 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
book-mirror
Take any book (EPUB/PDF), produce a personalized chapter-by-chapter analysis. Each chapter is preserved in detail (The Chapter) and mirrored back to the reader's actual life (The Mirror) using brain context. The mirror observes and resonates — a friend pointing out parallels, NOT a consultant rearranging the reader's…
ljg-learn
Deep concept anatomist that deconstructs any concept through 8 exploration dimensions (history, dialectics, phenomenology, linguistics, formalization, existentialism, aesthetics, meta-philosophy) and compresses insights into an epiphany. Use when user asks to explain, dissect, or deeply understand a concept, term, or…
eli5
Explain research, papers, or technical ideas in plain English with minimal jargon, concrete analogies, and clear takeaways. Use when the user says "ELI5 this", asks for a simple explanation of a paper or research result, wants jargon removed, or asks what something technically dense actually means.
deck-course-module
暖纸背景 + Playfair, 左侧学习目标常驻, 含 MCQ 自测页.
pedagogy-review
Holistic pedagogical review of a lecture deck (.qmd or .tex). Checks narrative arc, prerequisite assumptions, worked examples, notation clarity, and deck-level pacing. Use when user says "pedagogy review", "does this teach well?", "is the flow right?", "will students follow?", "review the narrative", or before…
master-yinguang
Use when user asks about 印光大师, 净土, 念佛, 持名念佛, 十念法, 摄耳谛听, 老实念佛, 信愿行, 带业往生, 仗佛慈力, 自力他力, 竖出横超, 往生, 极乐, 阿弥陀佛, 净土三经, 敦伦尽分, 闲邪存诚, 因果报应, 文钞, 一函遍复, or wants teaching in 印光大师 Yinguang's voice. Triggers include "印光"、"文钞"、"老实念佛"、"信愿行"、"带业往生"、"仗佛慈力"、"横超竖出"、"都摄六根"、"净念相继"、"敦伦尽分"、"闲邪存诚"、"因果"、"十念法"、"摄耳谛听"、"一函遍复"、"净土三经"、"往生" — invoke…