hol-guard
Use when setting up HOL Guard, protecting local AI harnesses, reviewing Guard approvals or receipts, scanning Codex plugins, skills, MCP servers, marketplace packages, or running plugin-scanner verification before release.
- License 许可
- Apache-2.0
- 最近更新
- 2026-11-14
- Stars
- ★ 1.6k
- Forks
- 491
01 前言 Front Matter
- name
- hol-guard
- license
- Apache-2.0
- description
- Use when setting up HOL Guard, protecting local AI harnesses, reviewing Guard approvals or receipts, scanning Codex plugins, skills, MCP servers, marketplace packages, or running plugin-scanner verification before release.
03 SKILL.md 正文
HOL Guard
HOL Guard protects local AI harnesses before tools run. Use this skill when the user wants AI antivirus behavior, local approval review, Codex protection, Claude Code protection, MCP safety checks, skill/package verification, or release gates from hol-guard and plugin-scanner.
Hard Rules
- Never read
.envfiles. - Never bypass Guard approvals.
- Do not mark a workspace protected until a Guard command proves status.
- Prefer reversible Guard commands over direct harness config edits.
- Do not mutate user-level harness config unless the
hol-guardcommand owns that mutation. - Treat scanner failures as real until inspected.
- Preserve existing user changes and inspect
git status --shortbefore edits in a repo.
Install Check
Check both CLIs independently:
command -v hol-guard
command -v plugin-scanner
If hol-guard is missing and the user asked for runtime setup, prefer:
pipx install hol-guard
If plugin-scanner is missing and the user asked for scanning, install the separate scanner distribution:
pipx install plugin-scanner
Do not assume the hol-guard distribution provides the plugin-scanner command. If pipx is unavailable, explain that isolated CLI installation is recommended rather than silently changing the user's Python environment.
After runtime installation:
hol-guard status
hol-guard detect --json
Protect A Local Harness
Use this flow for Codex, Claude Code, Copilot CLI, Cursor, Gemini, Hermes, OpenClaw, OpenCode, or Antigravity.
hol-guard bootstrap
hol-guard install <harness>
hol-guard run <harness> --dry-run
hol-guard run <harness>
hol-guard status
Harness names:
codexclaude-codecopilotcursorgeminihermesopenclawopencodeantigravity
Harness aliases:
claudemaps toclaude-codegemini-climaps togeminiopen-codemaps toopencodeopen-clawmaps toopenclawcopilot-climaps tocopilot
Use harness-specific bootstrap when available:
hol-guard hermes bootstrap
Claude Code
Use this when the workspace has .claude/settings.local.json, .claude/agents, Claude hooks, .mcp.json, or Claude-managed tool approval surfaces.
hol-guard install claude-code
hol-guard run claude-code --dry-run
hol-guard run claude-code
hol-guard doctor claude-code --json
Claude Code is a first-class Guard target. Prefer Guard-owned Claude hooks over direct manual edits to Claude config.
Codex
Use this when the workspace has Codex config, .codex/hooks.json, Codex MCP servers, or Codex App/CLI tool flows.
hol-guard install codex
hol-guard run codex --dry-run
hol-guard run codex
hol-guard doctor codex --json
Codex supports Guard-owned PreToolUse Bash hooks and same-chat MCP elicitation where available.
Other Harnesses
Use the same Guard flow for:
- Copilot CLI:
hol-guard install copilot - Cursor:
hol-guard install cursor - Gemini CLI:
hol-guard install gemini - Hermes:
hol-guard hermes bootstrap - OpenClaw:
hol-guard install openclaw - OpenCode:
hol-guard install opencode - Antigravity:
hol-guard install antigravity
Approval Work
If Guard blocks or queues work:
hol-guard approvals
hol-guard approvals open
hol-guard receipts
hol-guard diff <harness>
For terminal-only resolution:
hol-guard approvals approve <request-id>
hol-guard approvals deny <request-id>
Only approve after reading the risk reason and understanding the requested scope.
Evidence Work
Use evidence commands when user needs proof, audit trail, or handoff artifacts:
hol-guard receipts
hol-guard inventory
hol-guard abom --format json
hol-guard events
hol-guard explain <artifact-id>
For cloud sync, keep it optional and user-directed:
hol-guard connect
hol-guard connect status
hol-guard connect repair
hol-guard sync
Scan A Plugin Or Skill Package
Use scanner mode for Codex plugins, Claude Code project surfaces, .agents marketplaces, skills, MCP server configs, and release gates.
plugin-scanner lint .
plugin-scanner verify .
If scanning a specific package:
plugin-scanner lint <path>
plugin-scanner verify <path>
If the target is a Codex marketplace root with .agents/plugins/marketplace.json, scan the repo root so local plugin entries can be discovered.
Scanner target guidance:
- Codex plugin: scan the repo root or plugin folder containing
.codex-plugin/plugin.json. - Codex marketplace: scan the repo root containing
.agents/plugins/marketplace.json. - Claude Code project: scan the workspace root containing
.claude/,.mcp.json, hooks, or agent folders. - MCP server package: scan the package root containing server config and package metadata.
- Skill package: scan the folder containing
SKILL.md. - Mixed agent workspace: scan the repo root so local plugin, skill, MCP, and harness config surfaces are discovered together.
Common Debug Commands
hol-guard doctor
hol-guard doctor <harness> --json
hol-guard detect --json
hol-guard settings show
hol-guard explain install-connect
plugin-scanner verify . --json
Response Pattern
When using Guard, report:
- What command ran.
- What Guard found.
- What remains blocked or risky.
- What proof exists.
- Exact next command if user must act.
Do not claim protection, approval, or release readiness without command output proving it.
04 同仓库更多 Skills
worktable-review-with-annotations
Create, inspect, reply to, update, or resolve situated Worktable annotations. Use when feedback belongs to a Doc section, block, text range, or HTML Doc, or when the user asks to address existing review feedback and coordinate a requested artifact edit.
worktable-find-and-synthesize-context
Find, assess, and synthesize context from Worktable Docs and Records. Use when the user asks what a workspace contains, wants workspace-backed research, needs current facts reconciled across artifacts, or requests a cited synthesis without changing workspace state.
worktable-create-or-update-docs
Create, edit, rename, or maintain narrative Worktable Docs. Use for plans, decisions, briefs, notes, research syntheses, and other durable prose when the user requests a new artifact or a safe update to an existing one.
worktable-create-or-manage-records
Organize repeated structured information as durable Worktable Records. Use for research datasets, inventories, directories, pipelines, trackers, and other collections whose items need identity, fields, filtering, links, validation, or individual updates.
worktable-create-or-manage-interactive-html
Create or improve durable visual and interactive artifacts in Worktable. Use for dashboards, trackers, calculators, mini apps, slide presentations, and other requests that benefit from a polished HTML Doc.
worktable-collaborate-threads
Read, answer, and hand off messages through durable Worktable threads with users or other authenticated participants. Use when the user asks to check addressed messages, reply to a participant, continue a thread, request another agent's help, wait briefly for a response, or watch for comments and messages sent to the agent.