claude

self-updating-skills

Self-updating Agent Skills framework and harness authoring suite (self-updating-skills, claude-code-authoring, codex-authoring, opencode-authoring) powered by a single-file upkeep.sh runtime for version pin checks, reference freshness, live doc fetching, and SHA-256 frozen owner-policy blocks.

2 on GitHubby KHA Entertainmentv0.1.0MITupdated 4 days ago
/

Plugin self-updating-skills from marketplace kha-marketplace in KHAEntertainment/marketplace. Commands read from the repo's manifest — 3 days ago.

What it installs

Skills (4)

  • claude-code-authoring

    Author and review Claude Code extensions against the live, current docs instead of stale training data or old bundled guides. Use when creating, editing, reviewing or debugging a Claude Code skill (SKILL.md, frontmatter, $ARGUMENTS, context fork), hook (settings.json hooks, hooks.json, exit codes, d

  • codex-authoring

    Author and review OpenAI Codex CLI and IDE extensions against the live Codex documentation instead of stale training data. Use when creating, editing, or debugging AGENTS.md instructions, .codex/config.toml or requirements.toml, Codex skills (.agents/skills/<name>/SKILL.md and agents/openai.yaml), c

  • opencode-authoring

    Author and review OpenCode extensions and configuration against the live OpenCode documentation instead of stale training data. Use when creating, editing, or debugging opencode.json / opencode.jsonc, custom agents (.opencode/agents/*.md or agent config), permissions, custom tools (.opencode/tools/*

  • self-updating-skills

    Design, scaffold, retrofit, stamp, and audit self-updating Agent Skills (SKILL.md) that stay accurate when upstream CLIs, libraries, APIs, or coding harnesses ship breaking changes. Use when creating a new skill for a moving target, retrofitting an existing skill with version pins or freshness check

View source on GitHub khaentertainment.com

More from KHAEntertainment/marketplace

All 9 plugins →

view-limits

claude

Check coding-plan credit and rate-limit status across providers before dispatching sub-agents.

10 skills · 4 commands · hooks

2

enigma

claude

Local-only MCP secret request/reveal server (CLI + Claude Code plugin): request a named secret via a one-time URL; the value is not returned in MCP tool responses, but child-process output can expose it. `enigma_reveal` discloses it to the human directly; `enigma run` injects it into a child process's environment (ADR-001). Repo-level project scope (0.3.0): secrets saved from a worktree, symlinked clone path, or submodule need `enigma migrate-scope` after upgrade.

1 skill · 6 commands · 1 MCP server · hooks

2

dev-skill

claude

Turns Claude Code into a Tech Lead running an Issue-to-PR SOP: requirements, worktrees, delegated workers, QA, review, and retro gates.

2 skills

2

authoring-kit

claude

Author Claude Code skills, hooks, subagents and plugins against the live code.claude.com docs instead of stale training data or bundled guides.

1 skill

2

jev

claude

Decide when and how to integrate TypeSafe's Jev decision model: access paths (TypeSafe, OpenRouter, Vercel, Cloudflare), framework choice, harness patterns (tool gates, routers, triage), calibration and production ops.

1 skill

2

npm-publish

claude

Current npm publishing methodology — OIDC trusted publishing, staged publishing, provenance, and 2FA enforcement, token-free. Overrides stale training data that defaults to `npm token create` and `NPM_TOKEN` secrets.

1 skill

2

Maintain self-updating-skills?

This page updates itself from your repo. Add the badge to your README, or get it Featured to pin it above the organic results with a followed link to your site.

claude & codex plugins9 plugins