Context & AI engineering
25,745 rows, by stacks then stars
1,865 of these skill files have been read, by 1,328 distinct authors, and what they tell an agent counted. What Context & AI engineering authors agree on, and what they forbid
DROOdotFOO/agent-skills/skills/recall Skill
1★ repoQuery the recall knowledge base for relevant context. Searches decisions, patterns, gotchas, links, and insights stored across sessions. TRIGGER when: user asks to "recall", "remember", "what did we decide about", "check knowledge base", or references past decisions/patterns. DO NOT TRIGGER when: user is talking about memory/recall in a general sense, or working on the recall agent code itself.
DROOdotFOO/agent-skills/skills/regen Skill
1★ repoRead Fluidify Regen incidents and extract the join keys needed to pivot from an incident into its underlying OTel telemetry. Regen is a self-hosted incident / on-call tool with a REST API but no MCP server; this agent makes incidents first-class and emits a ready-to-run SigNoz filter hint -- the incident side of the incident <-> telemetry loop (SigNoz is the telemetry side). TRIGGER when: user asks about Regen incidents, on-call/paging state, "what's on fire", incident triage, ack/resolve an incident, or invokes "/regen"; or wants the service.name + labels + time window to feed into SigNoz. DO NOT TRIGGER when: user is working on the regen agent code itself; or wants to query the telemetry directly (use the signoz skill -- regen hands off to it).
DROOdotFOO/agent-skills/skills/self-improving-agent Skill
1★ repoAuto-curate memory, promote recurring patterns, and extract reusable knowledge across sessions. TRIGGER when: user says "/si:", asks about memory management, pattern promotion, or wants to review accumulated learnings. DO NOT TRIGGER when: user wants general code review, documentation, or project planning.
DROOdotFOO/agent-skills/skills/skill-creator Skill
1★ repoInteractively scaffold new Claude Code skills with correct frontmatter, trigger clauses, sub-files, and linter compliance. TRIGGER when: user asks to create a new skill, scaffold a skill, add a skill to agent-skills, or says "new skill" or "skill creator". DO NOT TRIGGER when: user is editing an existing skill, writing CLAUDE.md instructions, or building an MCP server (use mcp-server-builder skill).
arcjet/arcjet-plugin/skills/arcjet Skill
1★ repoAdd Arcjet security protection to any code path — HTTP route handlers, API endpoints, AI agent tool calls, MCP servers, background jobs, and queue workers. Covers rate limiting, bot detection, email validation, prompt injection detection, sensitive information blocking, and abuse prevention. Works across Next.js, Express, Fastify, SvelteKit, Remix, Bun, Deno, NestJS, FastAPI, Flask, and non-HTTP contexts. Use this skill when the user wants to add security, rate limiting, bot protection, or abuse prevention to any part of their application — whether they say "protect my API," "rate limit tool calls," "block bots," "secure my endpoint," "add security to my MCP server," or "prevent abuse" without mentioning Arcjet specifically.
ajinkya-t/codebase-oracle/skills/oracle-eval Skill
1★ repoAnalyze oracle usage and performance. Reads the eval log and produces a summary of cache hit rates, confidence distribution, cost estimates, and most-queried flows. Use to assess how well the oracle is working.
ladyiceberg/openclaw-memory-quality/skills/memory-diagnose Skill
1★ repoDeep diagnosis of OpenClaw retrieval quality — identifies false positives, zombie entries, and gives configuration recommendations to fix weak embedding or FTS-only retrieval. Use this skill whenever the user asks about retrieval quality, wonders why memory seems off, wants to understand false positives or zombie entries, asks about embedding or memory configuration, or uses phrases like "memory diagnosis", "bad memories", "why is memory bad", "memory config", "记忆诊断", "检索质量", "假阳性", "记忆配置". Prefer this skill over generic troubleshooting when OpenClaw memory quality is the topic.
ladyiceberg/openclaw-memory-quality/skills/memory-promote Skill
1★ repoPre-promotion quality audit for OpenClaw short-term memory candidates — checks the top-N entries before Dreaming promotes them to MEMORY.md, detects deleted source files, low-value snippets (imports/comments/debug code), already-promoted duplicates, and false-positive signals. Optional LLM long-term value advisory. Use this skill whenever the user wants to audit memories before Dreaming runs, check promotion candidates, review what's about to be promoted, or uses phrases like "promotion audit", "before dreaming", "check before promote", "memory promotion", "晋升前检查", "晋升审计", "Dreaming 前", "记忆晋升". Use this skill proactively when the user asks about Dreaming or upcoming memory promotions.
ladyiceberg/openclaw-memory-quality/skills/soul-check Skill
1★ repoAudit SOUL.md health for OpenClaw agents — detects boundary violations (injected code, URLs, shell commands, prompt injection), identity drift (action verb density, missing sections), and stability changes across runs. Optional LLM semantic conflict detection available. Use this skill whenever the user wants to check SOUL.md, mentions agent identity issues, soul drift, soul conflicts, or uses phrases like "check soul", "soul health", "SOUL.md", "agent identity", "soul drift", "soul conflict", "检查 SOUL", "SOUL 健康", "身份漂移", "语义冲突". Always use this skill for SOUL.md questions.
adrojis/tracecat-skills/skills/tracecat-mcp-tools-expert Skill
1★ repoActivate when users interact with Tracecat MCP tools, create workflows, manage cases, or need help with Tracecat API operations
adrojis/tracecat-skills/skills/tracecat-repo-sync Skill
1★ repoActivate when user asks to sync, push, or publish local Tracecat MCP server code or Claude Code skills to their GitHub repositories
adrojis/tracecat-skills/skills/tracecat-updates-monitor Skill
1★ repoMonitor Tracecat GitHub releases, documentation changes, API updates, and MCP developments. Run at session start or on-demand to detect new versions, breaking changes, and features that may impact our MCP server or skills.
varunk130/claude-code-skills/plugin-skills/document-skills/mcp-builder Skill
1★ repoGuide for creating high-quality MCP (Model Context Protocol) servers that enable LLMs to interact with external services through well-designed tools.
varunk130/claude-code-skills/plugin-skills/document-skills/skill-creator Skill
1★ repoGuide for creating effective skills that extend Claude's capabilities with specialized knowledge, workflows, or tool integrations.
varunk130/claude-code-skills/skills/financial-analysis/ai-agent-financial-analyst Skill
1★ repoSaaS financial modeling engine that generates unit economics models, feature ROI calculators, pricing scenario analyses, TAM/SAM/SOM sizing, build-vs-buy comparisons, and revenue projections from natural language inputs. Use when building business cases, calculating LTV/CAC/payback, modeling pricing changes, estimating feature revenue impact, running sensitivity analyses, or preparing financial justifications for product investments. Produces actual calculations with explicit assumptions and sensitivity ranges.
nevenincs/vaultspec-core/.vaultspec/skills/vaultspec-rag-discovery Skill
1★ repoSemantic codebase and architecture-decision discovery with vaultspec-rag - find code and the ADRs that govern it by meaning, then narrow with advanced filters and noise controls. Use to locate where or how something is done, or the decision behind it, instead of guessing identifiers or sweeping with keyword/grep search.
varunk130/AI-Eval-Skills/skills/eval-result-interpreter Skill
1★ repoAnalyzes AI agent evaluation results - primarily from Copilot Studio (the worked example here, via its CSV export) but also from custom harnesses or any evaluator that produces per-case pass/fail rows - using Microsoft's Triage & Improvement Playbook. Returns a SHIP / ITERATE / BLOCK verdict with root cause classification, diagnostic triage, prioritized remediation, and pattern analysis.
varunk130/AI-Eval-Skills/skills/eval-suite-planner Skill
1★ repoProduces a concrete eval suite plan for AI agents - grounded in Microsoft's Eval Scenario Library and MS Learn agent evaluation guidance (Copilot Studio is the primary worked example, but the plan is platform-agnostic and adapts to any agent harness). Outputs scenario types, evaluation methods, quality signals, thresholds, and priority order - before any test cases are generated or evals are run.
varunk130/AI-Eval-Skills/skills/eval-triage-and-improvement Skill
1★ repoUse this skill when AI agent evaluations have come back and the user needs to interpret scores, diagnose root causes of underperforming test cases, find remediation steps, or analyze patterns to improve their agent. Works against any agent platform - Copilot Studio is the primary worked example here, but the triage framework applies equally to custom harnesses, LangChain/LangGraph, AutoGen, Semantic Kernel, OpenAI Assistants, and other agent runtimes. Always use this skill when the user mentions: "eval failed", "why did this fail", "triage", "diagnose failure", "low pass rate", "fix evaluation results", "not passing", "failing test cases", "evaluation results", "improve my eval scores", or any situation where eval scores need interpretation and action.
varunk130/AI-Eval-Skills/skills/tool-use-eval Skill
1★ repoEvaluation pattern for agents that call tools / functions - measures tool selection accuracy, argument correctness, recovery from tool errors, and avoidance of unnecessary tool calls. Pairs with eval-suite-planner. Use when: tool-use eval, function-calling eval, agent tool selection, tool invocation accuracy, tool argument validation, tool error recovery, agent tools, function-calling agent, MCP eval.
Lu1sDV/skillsmd/llm-domain-speedrun Skill
no license1★ repoUse when someone needs to rapidly learn an unfamiliar technical domain under a tight deadline, when previous attempts through books, courses, or unstructured practice have failed, or when bridging a fundamental knowledge gap using an LLM as a private tutor
Lu1sDV/skillsmd/sink-research-orchestrator Skill
no license1★ repoUse when sink subagents encounter unseen languages or new technique categories that require fresh comprehensive research — not well-known sinks. Orchestrates parallel research swarms with structured JSON citations. Triggers: "comprehensive", "exhaustive", "most complete", "deep dive", "unseen language", "new technique category", "long horizon", "parallel research".
Universal research orchestrator
Lu1sDV/skillsmd/universal-research-orchestrator Skill
no license1★ repoUse when starting research, audit, or investigation tasks needing multi-source coverage — security/sink research, codebase audits, framework deep-dives, market comparison, or any "exhaustive / most complete / long horizon" request. Triggers on "audit", "investigate", "research", "find vulns", "exhaustive", "long horizon", "deep dive", "most complete".
Lu1sDV/skillsmd/zero-dof Skill
no license1★ repoUse when directing LLM coding agents on substantial development tasks — establishing executable oracles, designing agent playbooks, constraining agent output quality, or preventing metric gaming. Also use when LLM-generated code has recurring quality or architecture issues that need systematic correction. Triggers: zero dof, zero degrees of freedom, executable oracle, constrain agent, agent quality, LLM coding constraints, playbook design, oracle pipeline.
MARUCIE/openclaw-foundry/web/public/packs/spellbook-ai-app-engineer/skills/complex-doc-rag Skill
1★ repoUse when building a RAG pipeline that ingests PDFs, Excel, CSV, or images — especially when debugging silent data loss, choosing between OCR tools, or handling edge cases like scanned pages, merged cells, or embedded charts.
MARUCIE/openclaw-foundry/web/public/packs/spellbook-ai-app-engineer/skills/memory-map Skill
1★ repoUse when installing or configuring memory_map, writing CLAUDE.md session-setup instructions, choosing what to save in memory vs history, managing cross-project memory, tuning compression, or troubleshooting why Claude isn't loading context at session start.
MARUCIE/openclaw-foundry/web/public/packs/spellbook-code-reviewer/skills/complex-doc-rag Skill
1★ repoUse when building a RAG pipeline that ingests PDFs, Excel, CSV, or images — especially when debugging silent data loss, choosing between OCR tools, or handling edge cases like scanned pages, merged cells, or embedded charts.
MARUCIE/openclaw-foundry/web/public/packs/strategy-roundtable-advisor/skills/product-management-swarm Skill
1★ repo> SWARM (parallel experts, dynamic selection) | 产品战略、用户研究、市场分析、功能优先级、商业模型、竞品分析
yoprotocol/yo-protocol-skills/skills/yo-base-mcp-plugin Skill
1★ repoBase MCP plugin for YO Protocol — ERC-4626 yield vaults with ERC-7540 async redemption on Ethereum (1), Base (8453), and Arbitrum (42161). Both deposit and withdraw route through the YO Gateway. Use when the user wants to view YO vaults, check positions/balances/rewards, deposit, request a redeem (instant or async), or claim Merkl rewards via Base MCP's `send_calls`. Triggers on YO Protocol, yoUSD, yoETH, yoBTC, yoEUR, yoGOLD, yoUSDT, "deposit into yo", "withdraw from yo", "yo vault", or "claim yo rewards".
kxwu222/SEO-AEO-GEO-Assistant/skills/geo-visibility Skill
1★ repoGEO and AI-visibility Skill focused on getting content cited and quoted in AI-generated answers across Google AI Overviews, ChatGPT, Perplexity, Claude, Gemini, and Copilot.
0xc000022070/agentic-flake/tests/fixtures/external-repo/novu-skills/skills/connect-agent Skill
no license1★ repoFixture for example-4 checks.
anotherben/claude-enterprise-skills/skills/enterprise Skill
1★ repoEnterprise-grade development from idea to shipped product. Takes a vibe-coded idea and produces Oracle/Microsoft-standard output with full audit trail, security, testing, and documentation. Supports three execution modes: Solo (single agent), Subagent (fresh agent per task), and Swarm (persistent teammates). Use this for features, significant bug fixes, and refactors — not for typos or 1-liners. This is the primary development workflow.
The same file is in 2 other repositories
thedevappsecguy/skill-scanner/tests/fixtures/prompt_inject_skill Skill
1★ repoExample risky skill.
manastalukdar/ai-devstudio/skills/agent-evaluate Skill
1★ repoDefine behavioral contracts, run adversarial tests, and detect regressions for AI agents — invariants, edge cases, statistical analysis, and benchmark-production gap detection
manastalukdar/ai-devstudio/skills/context-budget Skill
1★ repoReport estimated context window utilization and flag when approaching the compaction threshold
manastalukdar/ai-devstudio/skills/context-engineering Skill
1★ repoAudit and curate the 5-layer agent context hierarchy to prevent quality degradation mid-session — rules files, specs, source files, error output, and conversation history.
manastalukdar/ai-devstudio/skills/database-connect Skill
1★ repoDatabase MCP server integration for PostgreSQL, MySQL, MongoDB
manastalukdar/ai-devstudio/skills/git-worktree Skill
1★ repoManage git worktrees for parallel feature development
manastalukdar/ai-devstudio/skills/llm-qa Skill
1★ repoInvestigate LLM output quality issues — trace hallucination root causes across context, retrieval, prompts, tool use, and workflow design
manastalukdar/ai-devstudio/skills/mcp-setup Skill
1★ repoSet up and configure MCP servers
manastalukdar/ai-devstudio/skills/memory-leak Skill
1★ repoMemory leak detection and analysis for Node.js, Python, and browsers
manastalukdar/ai-devstudio/skills/parallel-agents Skill
1★ repoManage multiple Claude instances for complex tasks with coordination
manastalukdar/ai-devstudio/skills/prompt-engineer Skill
1★ repoDesign, optimize, and evaluate prompts for LLM applications — system prompts, few-shot examples, chain-of-thought, structured outputs, and caching strategy
manastalukdar/ai-devstudio/skills/prompt-injection-detect Skill
1★ repoScan skill files, CLAUDE.md, MCP configs, and agent prompts for prompt injection patterns
manastalukdar/ai-devstudio/skills/rag-implement Skill
1★ repoDesign and implement Retrieval-Augmented Generation systems — chunking strategy, embedding selection, vector store setup, retrieval pipeline, re-ranking, and evaluation
hoveychen/claude-skills/plugins/cli-with-claude-skill/skills/cli-with-claude-skill Skill
1★ repoUse when creating, scaffolding, or adding/renaming/removing subcommands or flags in a CLI tool that ships an installable Claude Code skill (e.g. `<tool> install-claude-skill`), or when asked to "update the <tool> skill" / "the skill is out of date" for such a tool
hparamore/essential-claude-skills-hparamore/fable-toolkit/skills/session-handoff Skill
no license1★ repoWrite a clear, complete internal session-handoff / work-status log entry so the next session — you, the user, or another developer — can resume instantly with zero context loss. Use whenever the user says "update work status", "wrap up", "write a handoff", "log what we did", "save progress", "note where we left off", or reaches the natural end of a work chunk. Captures what changed, WHY, what's connected, gotchas, and next steps in plain language a stranger could follow. Not for user-facing release notes, a git changelog, or API/code documentation — this is the internal running log of a work session for whoever picks it up next.
Methasit-Pun/data_engineer_claude_skills/05-etl-build/orchestration-patterns Skill
no license1★ repoAirflow/Prefect/Dagster DAG design — task dependencies, retries, SLAs, backfill strategies, sensors, and failure recovery. Use this skill whenever the user is building or debugging a scheduled pipeline with multiple steps, asking how to handle task failures, setting up retries or alerts, designing a DAG structure, choosing between orchestrators, or dealing with backfill/reprocessing of historical data. Also trigger when the user mentions Airflow operators, Prefect flows, Dagster assets, task queues, or pipeline scheduling — even if they don't say "orchestration" explicitly. If a pipeline has more than two steps and needs to run on a schedule, this skill should be active.
Methasit-Pun/data_engineer_claude_skills/_tooling/agent-sort Skill
no license1★ repoBuild an evidence-backed ECC install plan for a specific repo by sorting skills, commands, rules, hooks, and extras into DAILY vs LIBRARY buckets using parallel repo-aware review passes. Use when ECC should be trimmed to what a project actually needs instead of loading the full bundle.
The same file is in 3 other repositories
vinsonconsulting/claude-skill-foundry/skills/tui/bubbletea Skill
1★ repoUse this skill when building a terminal UI (TUI) in Go with Bubble Tea and the Charm stack: the Elm Architecture / MVU pattern (`tea.Model` with `Init`/`Update`/`View`), `Cmd`/`Msg` event flow, Lip Gloss styling and layout, Bubbles components (`viewport`, `list`, `table`, `textinput`, `textarea`, `spinner`, `progress`), Glamour markdown rendering, and teatest for testing. Especially apt for streaming tokens from an LLM/agent into a terminal (goroutine → `p.Send` → `Update` → `viewport`), an agentic CLI, or any interactive full-screen or inline terminal app in Go. Targets v2 on the `charm.land/*/v2` import paths (Go 1.25+); writes current-version-correct code and avoids v1/beta patterns. Not for: Rust TUIs (→ Ratatui sibling), Python TUIs (→ Textual sibling), plain non-interactive CLI output (use `fmt`/`cobra`/standalone `lipgloss`), web/GUI UIs, or agent session orchestration / tmux / process multiplexing.
vinsonconsulting/claude-skill-foundry/skills/tui/ratatui Skill
1★ repoUse this skill for any task involving a Rust terminal/text user interface (TUI) built with Ratatui (or crossterm), whether creating one or debugging, fixing, testing, or extending an existing one. Covers: scaffolding the initial `main`/render loop and terminal setup; fixing teardown so a panic or crash doesn't leave the terminal in raw mode / broken; laying out panels, sidebars, status bars, dashboards, and popups with Layout/Constraint; wiring widgets like List, Table, Gauge, Chart, Scrollbar (including selection/scroll state that won't move); styling, text wrapping, and Unicode width issues; streaming LLM/async output into a terminal; and unit-testing rendered output without a real terminal. Triggers on "rust + terminal app/UI/dashboard", ratatui, crossterm. Writes current Ratatui 0.30+ code, not stale tui-rs patterns. Not for: Go TUIs (Bubble Tea), Python TUIs (Textual), non-interactive CLI output or progress bars, web/browser UIs, image-to-ASCII art (ascii/textmode), or tmux/agent-session orchestration.
vinsonconsulting/claude-skill-foundry/skills/tui/textual Skill
1★ repoUse this skill when building or debugging a Python terminal UI (TUI) with Textual (Textualize's framework): `App`/`Screen`/`Widget`, `compose()`, `reactive`/`watch_`, `@work` workers, Textual CSS / TCSS (`dock`, `grid`, `fr` units), widgets like `RichLog`, `Markdown`, `DataTable`, `Input`, `TextArea`, `Tree`, `ListView`; testing with `Pilot` / `pytest-textual-snapshot`; `textual serve` / Textual Web deploy; Rich renderables inside Textual; or streaming LLM/agent tokens into a terminal and agentic-CLI front-ends. Triggers: Textual, Textualize, TUI, terminal UI, `@work`, TCSS, `RichLog`, `query_one`, Pilot, `textual serve`. Writes current 8.x code, not stale pre-1.0 patterns. Not for: Rust TUIs (→ Ratatui sibling), Go TUIs (→ Bubble Tea sibling), plain non-interactive CLI output (Click/argparse/Typer) with no live UI, or agent-session orchestration / tmux multiplexing. Note: browser deploy via `textual serve` (the same Textual app served to a browser) IS in scope; a general React/HTML web-app request is not.
vikast908/agent-repo-card/skills/agent-reliability Skill
1★ repoUse when the user wants to know whether an AI agent / tool-using loop will survive the real world — reviewing loop termination, tool error handling, retries/backoff, idempotency, timeouts, state & resumability, guardrails, determinism, rate limits, graceful degradation, and observability/tracing. Triggers on "is my agent reliable", "review the agent loop", "why does the agent hang/loop forever", "production-readiness of my agent".
vikast908/agent-repo-card/skills/agent-security Skill
1★ repoUse when the user wants a security review of an AI/agent/LLM app — prompt injection, secret handling, tool permission scoping, sandboxing, data exfiltration, SSRF via tools, unsafe output handling, over-broad agent autonomy, and the OWASP LLM Top 10. Triggers on "is my agent secure", "security review", "can this be prompt-injected", "review for vulnerabilities", "is it safe to give the agent these tools".
vikast908/agent-repo-card/skills/prompt-quality Skill
1★ repoUse when the user wants to review the craft of the prompts in an AI/agent repo (not their token cost or injection safety) — clarity, structure, system/developer/user role separation, contradictions, brittle string concatenation, output contracts, few-shot quality, edge-case handling, testability, and maintainability. Triggers on "review my prompts", "are my prompts good", "improve this prompt", "why is the model ignoring instructions", "prompt engineering review".
vikast908/agent-repo-card/skills/report-card Skill
1★ repoUse when the user wants ONE combined quality grade for an AI-agent / LLM-app repo instead of running each review separately — auto-detects which reviews apply, runs them, dedupes overlapping findings, and emits a single overall grade, a per-area scorecard, and a prioritized cross-cutting fix list. Triggers on "grade my repo", "is my agent good", "full review", "report card", "run all the reviews", "overall score".
vikast908/agent-repo-card/skills/token-efficiency Skill
1★ repoUse when the user wants to reduce LLM token usage, context-window pressure, or API cost in an AI/agent codebase without hurting quality — reviewing prompt construction, context assembly, chat-history retention, tool definitions, retrieval, caching, batching, and output verbosity. Triggers on "reduce token usage", "cut LLM costs", "we're burning tokens", "optimize context", "why is this so expensive".
wemwi/skill-library/global-mcp-framework Skill
no license1★ repoGenerisches Framework zum Erstellen von Custom-MCP-Servern als Cloudflare Worker mit @cloudflare/workers-oauth-provider (OAuth 2.1), stateless Streamable-HTTP und einer gemeinsamen Foundation als versionierte Git-Tag-Dependency. IMMER laden, sobald ein MCP-Server in diesem Stack erstellt, verdrahtet, auf eine neue Foundation gebumpt oder debuggt wird — auch wenn das Wort Skill nicht fällt. Trigger u.a.: Custom MCP-Server bauen, Cloudflare Worker MCP, workers-oauth-provider, OAUTH_KV, OAuth 2.1 inbound, /authorize Login, Provider-Wiring, stateless Transport, sessionIdGenerator, "Session terminated", "Server misconfigured", "Authorization failed" nach Consent, neuen Connector anlegen, Foundation-Tag bumpen, wrangler.jsonc für MCP, KV-Namespace MCP_OAUTH, Discovery-Check well-known. Gilt für jeden neuen Custom-MCP-Server in diesem Stack.
t0ddharris/compound-marketing/.claude/skills/agent-browser Skill
no license1★ repoBrowser automation CLI for AI agents. Use when the user needs to interact with websites, including navigating pages, filling forms, clicking buttons, taking screenshots, extracting data, testing web apps, or automating any browser task. Triggers include requests to 'open a website,' 'fill out a form,' 'click a button,' 'take a screenshot,' 'scrape data from a page,' 'test this web app,' 'login to a site,' 'automate browser actions,' or any task requiring programmatic web interaction.
iamjosuho/memOrb/skills/core/writing-memorb-skills Skill
no license1★ repoMeta skill for adding or changing any sub-skill in this vault. Governs naming, frontmatter, structure, the registration checklist, and testing. Triggers: add a skill, write a skill, modify a skill, turn this into a skill, we keep doing this by hand, 新增 skill, 寫一個 skill, 修改 skill, 固化流程, 這個流程以後常做.