菜单
Extracts reusable patterns from the current session. Covers corrections, error resolutions, repeated workflows, Rails conventions. Use at the end of a productive session or when asked to "learn from this session".
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Reviews the current branch diff for bugs, edge cases, performance regressions, and behavior interactions. Use when asked for a code review, "review my changes", "check this code", or "anything wrong with this". This is a quick, focused review — for comprehensive multi-agent reviews, use full-code-review instead.
Analyzes code against staff-engineer patterns (security, architecture, simplicity, completeness, hygiene). Use when asked for a staff or senior code review, "staff review", "pattern review", or "review this like a staff engineer". This is the most thorough single-agent review — for multi-agent reviews, use full-code-review.
Generate monthly client progress reports for RailsPilot. Uses gh CLI to list the user's merged PRs, applies feature sizing rules, and drafts a numbered-list email. Use when asked for a monthly progress report, client delivery summary, or "progress report for [Month]".
Quick-reference card for all ponytail modes, skills, and commands. One-shot display, not a persistent mode. Trigger: /ponytail-help, "ponytail help", "what ponytail commands", "how do I use ponytail".
Write a warm, human-to-human written response (handoffs, status updates, testing notes for QA, sign-off requests, confirming assumptions before merging, or escalating an open question). Use when the user asks to draft a response, reply to a thread, write a handoff, send testing notes, ask the team to confirm something, phrase a Slack or PR comment, or convert dev-flavored notes into a message a teammate would actually enjoy reading.
Trim editorial bloat from technical docs (integration writeups, ADRs, investigation notes, READMEs). Proposes a unified diff with per-cut rationale and waits for approval. Use when asked to "trim this doc", "tighten this writeup", "filter pass on X.md", "prune the doc", or "cut the slop from this writeup".
| name | rails-learn |
| description | Extracts reusable patterns from the current session. Covers corrections, error resolutions, repeated workflows, Rails conventions. Use at the end of a productive session or when asked to "learn from this session". |
Extract patterns from the current session and recent commits, then persist them into the appropriate place (patterns.md, CLAUDE.md, a skill, or lessons.md).
Gather context: read tasks/lessons.md,
.claude/skills/railspilot-staff-review/patterns.md, and .claude/CLAUDE.md
to see what is already captured. Skip anything already documented.
Scan recent work: git log --since="7 days ago" and skim the diffs of any
commits whose messages reference fixes, reverts, or "actually". Pair each
with the conversation context from this session.
Look for these signal types, ranked by value:
Categorize each candidate by destination:
tasks/lessons.md: corrections from this session (date, title, what
happened, correct approach, applies-to area).Present findings. For each candidate show the observed pattern with a concrete example, the proposed destination, and the draft text. Ask the user which to apply.
Apply approved changes:
tasks/lessons.md.tasks/lessons.md is promoted into patterns.md,
CLAUDE.md, or a skill, delete it from tasks/lessons.md. The promoted
destination is now the source of truth, so the file stays a short
queue of pending corrections rather than an unbounded log.