메뉴
Analyzes the current conversation to propose improvements to skills based on what worked, what failed, and edge cases discovered. Use when the user says "reflect", "improve skill", "learn from this", or at end of skill-heavy sessions.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
Audits the security posture of an AI agent (codebase or architecture description): finds toxic combinations, reviews MCP governance and guardrails, and produces a prioritized report with an optional red-team plan. Use when reviewing agent security, auditing MCP configs, or assessing prompt-injection and exfiltration risk.
Manage Plane.so project management via the planecli CLI — list, create, update, and search work items, projects, cycles/sprints, modules, labels, states, documents, and comments. ALWAYS use this skill when the user mentions Plane, Plane.so, planecli, or references work item identifiers like ABC-123, FE-234, MOB-45. Also invoke when the user asks to manage tasks, issues, sprints, backlogs, or project boards AND either explicitly mentions Plane or the project context indicates Plane is the tool in use. Do NOT use for other project management tools (Jira, Linear, Asana, Trello, Azure DevOps Boards).
Analyzes coupling between modules using the three-dimensional model (strength, distance, volatility) from "Balancing Coupling in Software Design". Use when asking "are these modules too coupled?", "show me dependencies", "analyze integration quality", "which modules should I decouple?", "coupling report", or evaluating architectural health. Do NOT use for domain boundary analysis (use domain-analysis) or component sizing (use component-identification-sizing).
Reviews frontend files (`.tsx`, `.ts`, `.js`) for code quality, performance, and business logic issues. Use when the user requests a code review of frontend files, when reviewing staged changes before commit, or when checking specific components for best-practice violations.
Generates Vitest + React Testing Library tests for frontend components, hooks, and utilities. Use when writing tests, reviewing spec files, improving coverage, or when the user mentions Vitest, RTL, unit tests, or integration tests.
Produce HTML output instead of markdown for any agent task. Four levels — Static Doc, Visual Artifact, Two-Way Interactive, Throwaway Tool. Based on Thariq Shihipar's "Unreasonable Effectiveness of HTML" framework. Triggers on "/html-it", "html-it", "html this", "render as html", "build me an html", "make an html artifact".
| name | improving-skills-from-sessions |
| description | Analyzes the current conversation to propose improvements to skills based on what worked, what failed, and edge cases discovered. Use when the user says "reflect", "improve skill", "learn from this", or at end of skill-heavy sessions. |
Analyze the current conversation and propose improvements to a skill based on corrections, successes, and edge cases.
Run after a session where a skill was used. Optionally pass a skill name: /skill-improve-from-session [skill-name]
Copy this checklist:
Skill Reflection Progress:
- [ ] Step 1: Identify the target skill
- [ ] Step 2: Analyze conversation signals
- [ ] Step 3: Propose changes
- [ ] Step 4: Apply or save observations
If skill name not provided, use AskUserQuestion to ask which skill to analyze. List skills found in .claude/skills/.
Scan the current conversation for these signals:
Corrections (HIGH confidence):
Successes (MEDIUM confidence):
Edge Cases (MEDIUM confidence):
Preferences (accumulate over sessions):
Present findings in a summary box:
┌─ Skill Reflection: [skill-name] ─────────────────────────┐
│ │
│ Signals: X corrections, Y successes, Z edge cases │
│ │
│ Proposed changes: │
│ │
│ [HIGH] + Add constraint: "[specific constraint]" │
│ [MED] + Add preference: "[specific preference]" │
│ [LOW] ~ Note for review: "[observation]" │
│ │
│ Commit: "[skill]: [summary of changes]" │
│ │
└───────────────────────────────────────────────────────────┘
Ask user: Apply these changes? [Y/n] or describe tweaks
If approved:
.claude/skills/[skill-name]/SKILL.mdIf declined:
Ask: "Save these observations for later review?"
If yes, append to .claude/skills/[skill-name]/OBSERVATIONS.md
User runs /skill-improve-from-session frontend-design after a UI session:
┌─ Skill Reflection: frontend-design ──────────────────────┐
│ │
│ Signals: 2 corrections, 3 successes │
│ │
│ [HIGH] + Constraints/NEVER: │
│ "Use gradients unless explicitly requested" │
│ │
│ [HIGH] + Color & Theme: │
│ "Dark backgrounds: use #000, not #1a1a1a" │
│ │
│ [MED] + Layout: │
│ "Prefer CSS Grid for card layouts" │
│ │
│ Commit: "frontend-design: no gradients, #000 dark" │
│ │
└───────────────────────────────────────────────────────────┘
.claude/skills/) not absolute paths