InnerVault

Library

85,762 resources from 11 public sources. 54 checked by hand; the rest are unreviewed.

Results

  1. agent-leash

    Skills

    LEASH-8: an 8-domain control model for AI agents with delegated authority. Scorecard, approval-design checklist, plan-vs-authorize pattern. Patterns from a live production agent operation, sanitized. Free, MIT.

    Indexed GitHub, tonydzi 5 ★MITUpdated 3 days ago

    git clone https://github.com/tonydzi/agent-leash.git
  2. Agnostic-AI

    Skills

    Your AI coding harness, ported to every client you use. Capture Claude Code (or Codex) rules, hooks, skills, subagents, commands and MCP servers once; apply them identically to Codex, Gemini CLI, Cursor and 16 more. Zero-dependency Node.

    Stale GitHub, ucsandman 24 ★MITUpdated 8 days ago

    git clone https://github.com/ucsandman/Agnostic-AI.git
  3. Web-first Hermes skills index and safety gate: find SKILL.md packs before you install them.

    Stale GitHub, frankxai 7 ★MITUpdated 12 days ago

    git clone https://github.com/frankxai/awesome-hermes-agent-skills.git
  4. oasis

    Skills

    The open-source coding agent for any model: terminal and self-hosted web app, Claude Code plugins, subagents, and work that is actually verified.

    Indexed GitHub, FAHTHIZ 5 ★Apache-2.0Updated 2 days ago

    git clone https://github.com/FAHTHIZ/oasis.git
  5. Pro Tailscale workflows for every device you own, run by your AI agent with a safety tier on every command. Community skill, not affiliated with Tailscale.

    Stale GitHub, cathrynlavery 11 ★MITUpdated 26 days ago

    git clone https://github.com/cathrynlavery/tailscale-superpowers.git
  6. little-canary

    Skills

    Detects prompt injection by its effect on a sacrificial canary model, not just pattern matching: untrusted input hits a powerless model first, a behavioral check reads the residue, and it returns block, flag, or pass before your primary model acts. Inbound preflight sensor, not a guarantee.

    Stale GitHub, hermes-labs-ai 39 ★Apache-2.0Updated 28 days ago

    git clone https://github.com/hermes-labs-ai/little-canary.git
  7. aizee

    Skills

    The policy layer for AI coding agents. aiZee governs what Cursor, Claude Code, Copilot, Windsurf, Devin & MCP agents may do — 139 enforcement gates, 29 personas, 245 skills, BMAD runtime, hybrid memory, tamper-evident audit. pip install aizee

    Indexed GitHub, m3taz-ahmed 6 ★MITUpdated 2 days ago

    git clone https://github.com/m3taz-ahmed/aizee.git
  8. Zero-LLM deterministic jailbreak regression benchmark: runs a repeatable battery of known-pattern attacks against an LLM endpoint and scores refusal/partial/compliance across runs. Catch when a model or prompt update silently got weaker. Single-turn, responsible-use. pip install hermes-jailbench

    Stale GitHub, hermes-labs-ai 12 ★Apache-2.0Updated 28 days ago

    git clone https://github.com/hermes-labs-ai/hermes-jailbench.git
  9. 28 installable agent skills for construction: estimating, architecture, rendering, engineering, scheduling, site, safety, commercial and BD. Claude Code plugin, plain SKILL.md, CC BY 4.0.

    Stale GitHub, constructelligence-lab 10 ★Custom licence (see repository)Updated 8 days ago

    git clone https://github.com/constructelligence-lab/construction-agent-skills.git
  10. Browser automation with evidence. An Agent Skill that lets an AI agent drive a real browser while every action passes a policy gate, every run leaves an auditable artifact, and the run compiles into a plain Playwright script that replays for free.

    Indexed GitHub, OMD-123 13 ★MITUpdated 3 days ago

    git clone https://github.com/OMD-123/browser-verified-skill.git
  11. Claude-Abo

    Skills

    Portable Claude Code setup template: 66 skills with provenance, cost-aware model routing, git safety hooks. Adapt, don't run as-is.

    Stale GitHub, Bank848 6 ★Custom licence (see repository)Updated 8 days ago

    git clone https://github.com/Bank848/Claude-Abo.git
  12. Debloat an Android TV over ADB with an AI agent: brand-by-brand developer-mode entries, twelve safety gates, every step reversible. No root. 让 AI 替你清理电视。

    Stale GitHub, sanshengai 6 ★MITUpdated 22 days ago

    git clone https://github.com/sanshengai/sansheng-tv-adb.git
  13. lullsense

    Skills

    Open-source, evidence-informed baby & toddler sleep consultant (Agent Skill) — helps from conversation alone, sharper with recent sleep history; constraint-aware, safety-first, never diagnoses.

    Stale GitHub, Kavender 7 ★Apache-2.0Updated 8 days ago

    git clone https://github.com/Kavender/lullsense.git
  14. graph

    Skills

    Turns any idea into a checked plan your AI must follow: reads you first, finds and safety-checks the best skills, one-tap approval, rules enforced in code, full logs. Claude Code, Codex, Cursor, OpenCode, OpenClaw.

    Stale GitHub, gwaghmar 15 ★MITUpdated 6 days ago

    git clone https://github.com/gwaghmar/graph.git
  15. Nineteen Claude Code plugins I run every day across a Python backend, a React web app, a SwiftUI client and an ML pipeline: SDLC and ML skills, per-stack reviewer subagents, blocking safety hooks, content and SEO agents, LSP and monitors. One marketplace instead of a copy per repo.

    Indexed GitHub, kulykivska 6 ★MITUpdated 4 days ago

    git clone https://github.com/kulykivska/claude-plugins.git
  16. Set up Claude Code hooks to block dangerous git commands (push, reset --hard, clean, branch -D, etc.) before they execute. Use when user wants to prevent destructive git operations, add git safety hooks, or block git push/reset in Claude Code.

    Indexed Agent skill collections on GitHub, mattpocock MIT1 note

Indexed entries aren't reviewed or tested by InnerVault. Details come straight from each source. Anything that can reach your accounts, files or browser can do damage, so read the source before you install it.