Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add hashgraph-online/hol-guard --skill guardgit clone --depth 1 https://github.com/hashgraph-online/hol-guardWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/hashgraph-online/hol-guard/guard)<a href="https://agentmods.dev/skills/hashgraph-online/hol-guard/guard"><img src="https://agentmods.dev/badge/skills/hashgraph-online/hol-guard/guard/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/hashgraph-online/hol-guard/guard"><img src="https://agentmods.dev/badge/skills/hashgraph-online/hol-guard/guard.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00027 | $0.00532 |
| Opus 5.5 | $0.00011 | $0.00213 |
| Sonnet 5.5 | $0.00005 | $0.00106 |
| Haiku 4.5 | $0.00003 | $0.00053 |
Grade A, and why
hol-guard-protection scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 22d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 55 lines — stays where its author put it; the contents beside it link to each section on GitHub.
HOL Guard Skill Guidance
Use this guidance when an AI agent is about to add or update dependencies.
Before package installs
- Preview the install decision:
hol-guard protect --dry-run -- npm install <package>
- Check current workspace risk posture:
hol-guard supply-chain scan --json
- Explain a specific package verdict:
hol-guard supply-chain explain <package>@<version> --ecosystem <ecosystem>
- Confirm package manager interception is installed:
hol-guard package-shims status --json
- Repair a missing or tampered package manager shim:
hol-guard package-shims repair --manager npm --json
Cursor editor and Cursor CLI
Cursor has two protection surfaces. Cursor editor covers MCP servers in .cursor/mcp.json. Cursor CLI covers the cursor-agent command path. Keep them distinct when activating, checking, repairing, or removing Guard.
- Connect Cursor editor protection:
hol-guard apps connect cursor --surface editor
- Connect Cursor CLI protection:
hol-guard apps connect cursor --surface cli
- Test a Cursor surface without changing config:
hol-guard apps test cursor --surface editorhol-guard apps test cursor --surface cli
- Repair only the stale surface named by Guard Cloud:
hol-guard apps repair cursor --surface editorhol-guard apps repair cursor --surface cli
- Remove protection only after confirmation:
hol-guard apps disconnect cursor --surface editor --confirm disconnect-cursorhol-guard apps disconnect cursor --surface cli --confirm disconnect-cursor
Guard owns trust checks, drift repair, redacted receipts, and Cloud sync. Cursor owns its native editor and CLI behavior. If a surface is missing, unsupported, or unavailable, report that state instead of inventing an install URL or fallback command.
During CI and automation
- Route dependency installs through Guard:
hol-guard protect -- npm ci
- Use workspace audits before release:
hol-guard supply-chain audit --json
What ships with it
60 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- adapter-contributions.md 9.9 KB
- adr/0001-mdm-managed-install-contract.md 2.1 KB
- adr/0002-self-protection-authority.md 2.8 KB
- adr/0003-desktop-shell-and-tray-ownership.md 4.7 KB
- adr/0004-extension-control-center-semantics.md 5.7 KB
- adr/0005-protection-center-semantics.md 3.8 KB
- adr/0006-native-rust-runtime-boundary.md 2.2 KB
- adr/0007-native-runtime-packaging.md 3.0 KB
- adr/0008-native-resident-protocol-and-admission.md 6.1 KB
- adr/0009-native-and-daemon-critical-failure.md 2.9 KB
- adr/0010-native-posttool-default-auto.md 2.2 KB
- adr/0011-extension-first-managed-controls.md 6.5 KB
- adr/0012-everyday-and-technical-presentation.md 3.5 KB
- adr/0012-python-hook-capability-cleanup.md 1.8 KB
- adr/0013-native-resident-retirement.md 3.1 KB
- adr/0014-native-catalog-read-model.md 15 KB
- all-harness-hook-review.md 6.2 KB
- approval-invariant.md 6.1 KB
- architecture.md 3.9 KB
- audits/rm011-rtm006-gap-report.md 6.2 KB
- blocked-requests.md 2.3 KB
- business-apps/account-disconnect.md 1.8 KB
- business-apps/account-lifecycle.md 2.4 KB
- business-apps/account-refresh.md 2.7 KB
- business-apps/attempt-journal.md 1.6 KB
- business-apps/budget-contract.md 1.8 KB
- business-apps/budget-reservations.md 2.6 KB
- business-apps/compatibility.md 3.7 KB
- business-apps/directory-preparation.md 4.2 KB
- business-apps/dispatch-outcome-receipts.md 2.9 KB
- business-apps/gmail-plain-input.md 3.8 KB
- business-apps/gmail-wire-input.md 3.8 KB
- business-apps/gog-command-extension.md 2.3 KB
- business-apps/google-identity-evidence.md 4.0 KB
- business-apps/google-workspace-pack.md 2.8 KB
- business-apps/gws-command-extension.md 4.4 KB
- business-apps/gws-gmail-command-input.md 3.3 KB
- business-apps/local-review-summary-ui.md 6.5 KB
- business-apps/local-review-summary.md 1.9 KB
- business-apps/native-policy-binding.md 3.0 KB
- business-apps/outbound-inspection.md 1.7 KB
- business-apps/owned-dispatch-claim.md 2.7 KB
- business-apps/owned-native-dispatch.md 2.1 KB
- business-apps/owned-send-transport.md 2.1 KB
- business-apps/policy-document-drafts.md 1.1 KB
- business-apps/private-review-input.md 3.3 KB
- business-apps/salesforce-sf-command-extension.md 3.0 KB
- business-apps/source-map.md 6.7 KB
- business-apps/worker-input.md 1.8 KB
- business-apps/worker-oauth.md 5.9 KB
- business-apps/worker-peer-identity.md 1.7 KB
- catalog-discovery-preparation.md 6.2 KB
- ci-gate-ownership.md 3.3 KB
- cline-local-protection-contract.md 6.4 KB
- cline-security-closure.md 6.1 KB
- cline-upstream-evidence.md 3.0 KB
- cloud-command-capability.md 3.3 KB
- codex-binding-diagnostics.md 3.5 KB
- command-activity-dashboard-ia.md 3.7 KB
- command-activity-privacy.md 6.0 KB
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 22d ago Changed · +5 lines · +27 tokens per session 5ddcc67c255e
- 1mo ago First seen · 50 lines · 0 tokens per session scan A f1751c981869
hol-guard-protection is a skill published in the GitHub repository hashgraph-online/hol-guard (956 stars, last pushed today), licensed Apache-2.0. It adds 27 tokens to every session and 532 once invoked, about $0.0001 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
crossval-harness
Orchestrate a static + dynamic, exploit-validated red-team of an AI agent — read the source to find candidate vulnerable paths, then run the dynamic skills to confirm or refute each one empirically. The arbiter of truth is whether the exploit works, not a model vote. Authorized testing of agents you own or are…
redteam-an-agent
The end-to-end methodology for red-teaming a specific AI agent — adaptively, exploit-validated, and honestly. Read THIS target's own code, stand up a disposable harness, and prove or refute each weakness through a real attacker-reachable entry point. This is the orchestration + discipline that makes a finding…
audit-agent-code
Security-audit an AI agent's OWN framework code for classic appsec vulnerabilities the LLM can't defend — path traversal, command injection, SSRF, fail-open security controls, missing/late auth, unsafe deserialization — especially at the untrusted-input boundaries (channels, file/media handlers, config & skill…
rt6-memory-poisoning
Red-team an AI agent's persistent memory for cross-session prompt injection ("memory poisoning") — does untrusted content the agent processes get written into long-term memory and re-fire in future sessions with no attacker present? Authorized testing of agents you own or are permitted to test.
rt9-multi-agent
Red-team a multi-agent system — can one agent (or content it relays) inject instructions into another, escalate privilege by hopping between agents, or turn an orchestrator/sub-agent handoff into a trust-laundering path? Authorized testing of systems you own or are permitted to test.
rt1-prompt-injection
Red-team an AI agent for prompt injection — does content the agent is asked to process (email, page, ticket, tool result) override its actual task? Authorized testing of agents you own or are permitted to test.