build-better-skills

Quickly open-source a local skill to GitHub (primary) and optionally clawhub.com. Workflow: slug pre-check, fork to opensourceskills, strip internal info, normalize SKILL.md, generate LICENSE/README, init git, push to GitHub with configurable token source, and optionally publish to clawhub. Use when the user says "open-source this skill", "把这个 skill 开源", "发到 github", "publish skill publicly", or "把本地 skill 发开源". Hard rules: never modify the original skill in place, never auto-add force/yes flags, never write tokens to git/remote/memory, and ask the user at decision points.

Query, install, update, and edit AI agent skills on any compatible Skill Hub (self-hosted, or any Hub implementing the documented API contract). Dual-channel: with a token it uses the authenticated API (full features including private skills); without a token it falls back to the unauthenticated channel (public skills only). Covers: list newly published skills, search by keyword / author / time / source, inspect version history, install or upgrade a specific version, and edit a skill's card metadata (display name, summary, tags, visibility, applicable position, etc.) via a safety-first GET -> diff -> backup -> PUT -> dual-channel verify -> auto-rollback flow. Trigger phrases include "what's new on the hub", "search for X skill", "install X", "update Y skill", "edit hub card info", "show skill version history", "skill-hub-query".

Install and operate the OpenClaw skill-sediment plugin extension (turns successful conversations into auto-generated SKILL.md files, then promotes them to skills/). Covers: install (with interactive agent selection), heal (self-repair: detect and fix whitelist/file loss, supplement missing validAgentId, auto-restart gateway), doctor (diagnostics), recover (restore after pod restart), status (sediment pool stats), config (current state), uninstall. Triggers on: sediment / auto-skill-sediment / install/port/clone sediment / sediment not working / sediment pool / uninstall sediment. Note: unrelated to the same-named Rust semantic-memory MCP (store/recall); does not cover OpenClaw installation itself nor publishing to Skill Hubs.

End-to-end guided workflow for building production-quality agent skills. Walks users through 4 stages: pre-flight checks → spec confirmation (3 musts) → coding quality acceptance (4 checks) → release prep (5 requirements). Use when the user says "create a skill", "build a skill", "write an agent skill", "skill-build-wizard", "做个 skill", "帮我创建一个 skill", "skill 全流程向导". Calls skill-creator for actual scaffolding and routes to glic-check / skill-release-audit / skill-release-plus / skill-regression for verification and release. Not for: just writing one quick skill file (use skill-creator directly); just publishing an already-finished skill (use skill-release-plus).

Regression testing framework for AgentSkills. Analyzes a target skill, runs script-layer assertions and AI-layer semantic scoring, and outputs a Markdown report. Supports two backends: OpenAI-compatible LLM (default, works anywhere) or OpenClaw cron-based agent trigger (auto-detected if openclaw CLI present). Use when the user asks to "test a skill", "regression test xxx skill", "run skill QA", or "audit my skill against test cases". Requires SR_LLM_API_KEY env var or interactive setup; supports .env files.

Pre-publish quality and safety auditor for AI agent skills (SKILL.md + scripts/ + references/ format used by Claude Code, Cursor, OpenAI Codex, GitHub Copilot, OpenClaw, ClawHub, and compatible SkillHub registries). Six static-check modules (no LLM, no network by default): (1) syntax and logic correctness, (2) feature completeness, (3) edge-case and error handling, (4) data safety (detects files written inside the skill dir that would be lost on update), (5) dependency declaration vs code, (6) SKILL.md documentation standards. Per-registry rule profiles via `--target`. Pure reporter — never edits your files, never publishes. Use when publishing a skill, modifying an existing skill, or diagnosing why a skill behaves unexpectedly — run it as the last machine-checkable gate before release. Trigger phrases: "skill release audit", "audit before publishing", "pre-release check", "release gate", "skill safety check", "发版前检查", "skill 发布检查", "审查这个 skill 能不能发版".

Multi-hub skill publisher. Sign → Audit (placeholder) → Pack (clean tar.gz with exclude rules) → Publish to clawhub.com, skillhub.cn (Tencent Cloud), or GitHub Releases, all in one command. Pluggable adapter framework — add your own hub via a user-hook script (subprocess + JSON envelope). Single source of truth: one exclude.json, one signing pass, one package, fan out to N targets. Default target: clawhub only. Use --target all for all three real hubs, or --target user-hook:./my-script.sh for custom destinations. Triggers on: publish skill, release skill, push to skill hub, ship skill to multiple hubs, 发布 skill, 上传到 clawhub, 发布到 skillhub.cn.

Cryptographically sign and verify AgentSkill directories using Ed25519 (RFC 8032). Publishers sign skills with a private key; recipients verify with a public key, proving the skill is unmodified AND authored by the trusted key holder. Pure Python — no pip install, no native dependencies. Use when you say sign skill, verify skill, ed25519 sign, verify skill authenticity, prove skill author, detect skill tampering, give skill a signature, 给 skill 签名, 验证 skill 是谁发的, 校验 skill 没被改过.

Systematic quality check for code, skills, configs, and documents. Two modes — GLIC for internal quality (4 dimensions: Grammar / Logic / Integrity / Containment) and UGLIC adding User Experience (5 dimensions: U + G + L + I + C). Each finding cites file:line; severity is tagged as ERR / WARN / INFO with explicit escalation rules (silent-failure = ERR, 3× repeated WARN → ERR, missing public-param doc = ERR). Use when the user says "GLIC check", "UGLIC check", "do a glic", "systematically review this", "audit my skill", "quality check this code", or any phrasing that asks for a multi-dimension review of code / skills / configs / docs.

Generic skill-quality auditor for any agent skill (Claude, OpenClaw, Cursor, etc.). Runs a 7-dimension static analysis (D1 process closure & idempotency, D2 tool/command conventions, D3 portability & defense, D4 skill usability, D5 security & op risk, D6 code & doc quality, D7 dependency & footprint) with explicit ERR / WARN severity, 115-point scoring (pass line 90 + zero ERR), and an opt-in `--fix` workflow that always backs up first. Two depths: L1 static (~2 min) and L2 dryRun (~5 min, read-only hub + reachability checks). Strict red lines — read-only by default, never executes the audited skill's writes. Use when the user asks to "audit a skill", "check skill quality", "is this skill ready to ship", "lint my skill", or runs this tool by name. Triggers also: "审计这个 Skill"、"检查 Skill 质量"、"Skill 能上线吗"、 "skill-deep-audit"、"审一下 xxx skill"。

Install skills from multiple skill hubs with one tool: clawhub.ai (default), skillhub.cn (SkillHub, China-optimized), skills.sh (npx skills, GitHub-based), the official anthropics/skills repo, and your own self-hosted "custom" hub. Routes to the right source based on how the user phrases the request, handles name collisions (rename/overwrite/abort), Anthropic source-available license gating, multi-skill repos, and path-traversal-safe extraction. Use when the user says "install skill X", "add skill X", "download skill X", "get X from clawhub", "get X from skillhub.cn", "use skills.sh to install X", "install the claude/anthropic X skill", or "install X from my hub". Also triggers in Chinese: "安装skill xxx"、"装一下xxx"、"下载skill xxx"、 "从 clawhub 装 xxx"、"从 skillhub.cn 装 xxx"、"用 skills.sh 装 xxx"、 "装 claude 的 xxx"、"从我的 hub 装 xxx"。

Generate a beautiful, deployable HTML introduction page for any AgentSkill. Reads USAGE.md (preferred) or SKILL.md, parses name, description, and feature sections, then renders a polished page with a hero section, feature cards, quick start, command examples, collapsible doc cards, side navigation, and a floating install button. Supports 4 themes: light (default), aurora (dark glass), techblue (debug-tool style), finance (dark blue + gold). Optionally deploys via a pluggable hook so re-runs update the same URL. Use when the user says "generate a skill intro page", "make a doc page for this skill", "publish this skill's docs", "skill-introduction". If the user does not specify a target skill, ask them for the skill name or directory path.