[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M74GBTDVAEB853YKF29GS462":3,"skill-same:01M74GBTDVAEB853YKF29GS462":106},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":15,"categories":16,"keywords":17,"publishedAt":18,"linkOnly":19,"markdownContent":20,"markdownBody":21,"frontMatter":22,"refName":28,"commitSha":29,"skillDirectory":30,"forksCount":13,"repositoryHtmlUrl":31,"repositoryHomepageUrl":6,"ownerHtmlUrl":32,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":33,"install":34,"resources":47},"01M74GBTDVAEB853YKF29GS462","audit-openclaw-security",null,"Audit and harden OpenClaw (Gateway + agents) security. Use when the user asks to audit\u002Fsecure\u002Fharden OpenClaw; when troubleshooting risky exposure (especially the Gateway web UI\u002Fcontrol plane on port 18789); when reviewing DM\u002Fgroup access control (pairing\u002Fallowlists\u002Fmention-gating); tool permissions (exec\u002Ffs\u002Fbrowser\u002Fnodes\u002Fgateway\u002Fcron); plugins\u002Fskills supply-chain risk; secrets\u002Ftranscripts\u002Flog retention; or when deploying OpenClaw on a Mac mini, personal laptop, Docker host, or cloud VM (AWS EC2\u002FVPS).","meghal86\u002Fstacksignal","meghal86","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F37388715?v=4","01M6CS2EW8SK89GH52R2ZEN8QQ","01M728J5CGY456EETEFG6HJBTM",0,"MIT","https:\u002F\u002Fgithub.com\u002Fmeghal86\u002Fstacksignal\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Faudit-openclaw-security\u002FSKILL.md",[],[],"2026-11-11 15:11:14",false,"---\nname: audit-openclaw-security\ndescription: Audit and harden OpenClaw (Gateway + agents) security. Use when the user asks to audit\u002Fsecure\u002Fharden OpenClaw; when troubleshooting risky exposure (especially the Gateway web UI\u002Fcontrol plane on port 18789); when reviewing DM\u002Fgroup access control (pairing\u002Fallowlists\u002Fmention-gating); tool permissions (exec\u002Ffs\u002Fbrowser\u002Fnodes\u002Fgateway\u002Fcron); plugins\u002Fskills supply-chain risk; secrets\u002Ftranscripts\u002Flog retention; or when deploying OpenClaw on a Mac mini, personal laptop, Docker host, or cloud VM (AWS EC2\u002FVPS).\nlicense: MIT\ncompatibility: Works best in Claude Code (local shell access). In Claude.ai, instruct the user to run the included scripts\u002Fcommands locally and share redacted outputs (prefer `openclaw status --all` and `openclaw security audit --json`).\nmetadata:\n  author: \"community\"\n  version: \"2.0.0\"\n  upstream: \"OpenClaw docs + security guidance (updated Feb 2026)\"\n---\n\n# audit-openclaw-security\n\nTeach the agent to perform a **defensive, permissioned** security audit of an OpenClaw deployment and produce an actionable report + remediation plan.\n\nOpenClaw is a local-first personal AI assistant with a Gateway control plane, multi-channel inbox, tools (browser\u002Ffs\u002Fexec\u002Fnodes), and optional remote access patterns. The goal of this skill is to **reduce attack surface** (network + identity), **minimise agent permissions**, and **protect secrets\u002Ftranscripts**.\n\n## Scope and safety rules (non-negotiable)\n\n1. **Only audit systems you own or have explicit permission to test.**\n2. **Do not request or handle raw secrets.** Never ask for:\n   - gateway tokens\u002Fpasswords\n   - model API keys \u002F OAuth tokens\n   - WhatsApp creds, Slack tokens, Discord tokens, etc.\n3. Prefer outputs that are designed to be shareable\u002Fredacted:\n   - `openclaw status --all`\n   - `openclaw security audit --json` (and `--deep --json`)\n4. Treat **any remote browser control \u002F node pairing** as *admin-equivalent access* and audit accordingly.\n5. Default to **audit-only**. If the user wants remediation, propose changes and get explicit approval before running any:\n   - `openclaw security audit --fix`\n   - config edits (`openclaw config set`, Control UI Config, `config.apply` \u002F `config.patch`)\n   - firewall edits\n   - restarts\n\n## What “good” looks like (target posture)\n\n- Gateway binds to loopback unless there is a strong reason otherwise.\n- Strong Gateway auth (token\u002Fpassword, rotated) is enabled.\n- No public internet exposure (no open security group \u002F port-forward \u002F Tailscale Funnel \u002F accidental reverse proxy bypass).\n- DMs require pairing or strict allowlists; group chats require mention-gating.\n- Tooling is least-privilege (deny\u002Fask for exec; workspace-only FS; deny control-plane tools on untrusted surfaces).\n- Secrets are stored with tight filesystem permissions; logs\u002Ftranscripts have a retention plan.\n- Plugins\u002Fskills are explicitly trusted, minimal, and protected from unauthorised modification.\n\n## Workflow\n\n### Step 0 — Establish context (fast)\n\nCollect just enough context to pick the right audit path:\n\n- Where is OpenClaw running?\n  - **Mac mini** (always-on home host)\n  - **Personal laptop** (macOS\u002FWindows\u002FLinux)\n  - **Docker** (local or on a VM)\n  - **Cloud host** (AWS EC2 \u002F VPS)\n- Install type?\n  - native install vs Docker Compose vs source checkout\n- Do we have **local shell access**?\n  - **Claude Code**: can run commands directly\n  - **Claude.ai**: user must run commands and paste outputs\n\nThen pick one of these modes:\n\n- **Mode A: Assisted self-audit (Claude.ai)** — user runs commands\u002Fscripts, shares redacted output.\n- **Mode B: Automated local audit (Claude Code)** — run scripts directly and generate a report.\n\n---\n\n## Mode A — Assisted self-audit (no local shell access)\n\nAsk the user to run the following on the OpenClaw host and paste results.\n\n> If they are worried about sharing output, reassure them that `openclaw status --all` and `openclaw security audit --json` are intended to be shareable. If they still refuse, provide high-level hardening guidance without requesting outputs.\n\n### 1) Collect OpenClaw posture (recommended minimum)\n\n```bash\nopenclaw --version\nopenclaw status --all\nopenclaw doctor\nopenclaw security audit --json\nopenclaw security audit --deep --json\n```\n\nHelpful extras (for supply-chain + snapshot context):\n\n```bash\nopenclaw health --json\nopenclaw skills list --eligible --json\nopenclaw plugins list --json\n```\n\nIf the user is comfortable sharing a config for review, prefer **targeted** reads over full-file sharing:\n\n```bash\nopenclaw config get gateway.bind\nopenclaw config get gateway.auth.mode\nopenclaw config get discovery.mdns.mode\nopenclaw config get session.dmScope\n```\n\nIf they *must* share the full config, remind them OpenClaw config is **JSON5** (`~\u002F.openclaw\u002Fopenclaw.json`), not strict JSON, and have them use the redaction script:\n\n```bash\npython3 scripts\u002Fredact_openclaw_config.py ~\u002F.openclaw\u002Fopenclaw.json > openclaw.json5.redacted\n```\n\n### 2) Collect host\u002Fnetwork posture (pick the right block)\n\n**macOS (Mac mini \u002F Mac laptop):**\n```bash\nwhoami\nsw_vers\nuname -a\nlsof -nP -iTCP -sTCP:LISTEN\n\u002Fusr\u002Flibexec\u002FApplicationFirewall\u002Fsocketfilterfw --getglobalstate\n\u002Fusr\u002Flibexec\u002FApplicationFirewall\u002Fsocketfilterfw --getstealthmode\nfdesetup status || true\n```\n\n**Linux (EC2\u002FVPS):**\n```bash\nwhoami\ncat \u002Fetc\u002Fos-release\nuname -a\nss -ltnp\nsudo ufw status verbose || true\nsudo nft list ruleset || true\nsudo iptables -S || true\n```\n\n**Docker deployments (local or on a VM):**\n```bash\ndocker ps\n# confirm published listeners (should be 127.0.0.1:18789, not 0.0.0.0:18789)\ndocker port openclaw-gateway 18789 || true\ndocker compose ps || true\n```\n\n---\n\n## Mode B — Automated local audit (Claude Code)\n\nIf shell tools are available, run:\n\n```bash\nbash scripts\u002Fcollect_openclaw_audit.sh --out .\u002Fopenclaw-audit\npython3 scripts\u002Frender_report.py --input .\u002Fopenclaw-audit --output .\u002Fopenclaw-security-report.md\n```\n\nThen open `openclaw-security-report.md`, refine wording, and present the report to the user.\n\n---\n\n## Analysis and reporting\n\n### Step 1 — Triage findings by severity\n\nUse OpenClaw’s audit output as the source of truth. Prioritise in this order:\n\n1. **Anything “open” + tools enabled** (lock down DMs\u002Fgroups first; then tighten tools\u002Fsandbox)\n2. **Public network exposure** (LAN bind, Tailscale Funnel, reverse proxy mistakes, missing auth)\n3. **Browser control \u002F nodes exposure** (treat like operator access)\n4. **File permissions** on `~\u002F.openclaw` (config\u002Fstate\u002Fcredentials)\n5. **Plugins\u002Fextensions + dynamic skill loading** (supply chain)\n6. **Model hygiene + prompt-injection resilience** (important, but secondary to access control)\n\n### Step 2 — Translate findings into a clear risk narrative\n\nFor each finding, capture:\n\n- **What was observed** (quote `checkId` + summary; do not paste secrets)\n- **Why it matters**\n- **Impact** (data exfiltration, account takeover, unauthorised actions, persistence)\n- **Concrete remediation steps**\n- **Verification** (usually rerun `openclaw security audit --deep`)\n\n### Step 3 — Recommend a secure baseline configuration (when appropriate)\n\nOffer the baseline in `references\u002Fopenclaw-baseline-config.md` as a starting point and adapt based on needs:\n\n- single-user vs shared inbox\n- required tools\n- remote access method (SSH tunnel \u002F Tailscale Serve)\n\n### Step 4 — Platform-specific guidance\n\nUse the platform playbooks:\n\n- `references\u002Fplatform-mac-mini.md`\n- `references\u002Fplatform-personal-laptop.md`\n- `references\u002Fplatform-aws-ec2.md`\n- `references\u002Fplatform-docker.md`\n\n---\n\n## Deliverable\n\nProduce a Markdown security audit report using `assets\u002Freport-template.md`:\n\n- Executive summary (1–2 paragraphs)\n- Environment overview (host type, install type, exposure)\n- Findings table (Critical\u002FHigh\u002FMedium\u002FLow)\n- Remediation plan (sequenced)\n- Verification steps\n- Residual risk + operational practices (patching, token rotation, retention)\n\n---\n\n## Common remediation patterns (use with care)\n\n### Network exposure\n\n- Prefer Gateway bound to loopback and remote access via **SSH tunnel** or **Tailscale Serve**.\n- Avoid **Tailscale Funnel** for the Gateway.\n- Never expose the Gateway control UI directly to the public internet.\n- If you must run a reverse proxy, configure **trusted proxies** and block direct access to the Gateway port.\n\n### Identity & access\n\n- Use token\u002Fpassword auth; rotate periodically and after any suspected leakage.\n- If you need a token generated in automation: `openclaw doctor --generate-gateway-token`.\n- Require DM pairing or strict allowlists.\n- In groups, require mention gating; avoid always-on bots in large groups.\n\n### Control-plane tools\n\n- The `gateway` and `cron` tool surfaces can create persistence (config changes, scheduled jobs). Deny them on any surface that handles untrusted content.\n\n### Least privilege tools\n\n- Keep filesystem tools workspace-only.\n- Keep exec denied or “ask always” unless the user is present and approving actions.\n- Treat browser control like operator access; keep it tailnet\u002Flocal only.\n\n### Secrets, transcripts, and logs\n\n- Tight permissions on `~\u002F.openclaw` and `~\u002F.openclaw\u002Fopenclaw.json`.\n- Treat session transcripts as sensitive (`~\u002F.openclaw\u002Fagents\u002F\u003CagentId>\u002Fsessions\u002F*.jsonl`).\n- Gateway logs default under `\u002Ftmp\u002Fopenclaw\u002Fopenclaw-YYYY-MM-DD.log` (or `logging.file`).\n- Keep redaction enabled and set a retention\u002Fpruning policy.\n\n---\n\n## Troubleshooting\n\n### “openclaw: command not found”\n\n- Confirm OpenClaw CLI is installed and on PATH.\n- If using Windows, prefer WSL2.\n- Re-run installation via the official method for your platform; then re-run `openclaw --version`.\n\n### “refusing to bind ... without auth”\n\n- Expected guardrail: non-loopback binds require auth. Do not bypass.\n\n### Audit output is empty or inconsistent\n\n- Ensure the Gateway is running: `openclaw gateway status`\n- Re-run with `--deep` for best-effort live probing.\n\n---\n\n## Trigger tests (for skill author)\n\nShould trigger:\n\n- “Can you audit my OpenClaw setup for security?”\n- “My OpenClaw gateway is on port 18789 — is that safe?”\n- “I deployed OpenClaw on EC2; give me a hardening checklist.”\n- “Run openclaw security audit and interpret the results.”\n- “I’m running OpenClaw in Docker—how do I lock it down?”\n\nShould NOT trigger:\n\n- General macOS security advice unrelated to OpenClaw\n- Generic AWS hardening unrelated to OpenClaw\n- Unrelated software audits\n","# audit-openclaw-security\n\nTeach the agent to perform a **defensive, permissioned** security audit of an OpenClaw deployment and produce an actionable report + remediation plan.\n\nOpenClaw is a local-first personal AI assistant with a Gateway control plane, multi-channel inbox, tools (browser\u002Ffs\u002Fexec\u002Fnodes), and optional remote access patterns. The goal of this skill is to **reduce attack surface** (network + identity), **minimise agent permissions**, and **protect secrets\u002Ftranscripts**.\n\n## Scope and safety rules (non-negotiable)\n\n1. **Only audit systems you own or have explicit permission to test.**\n2. **Do not request or handle raw secrets.** Never ask for:\n   - gateway tokens\u002Fpasswords\n   - model API keys \u002F OAuth tokens\n   - WhatsApp creds, Slack tokens, Discord tokens, etc.\n3. Prefer outputs that are designed to be shareable\u002Fredacted:\n   - `openclaw status --all`\n   - `openclaw security audit --json` (and `--deep --json`)\n4. Treat **any remote browser control \u002F node pairing** as *admin-equivalent access* and audit accordingly.\n5. Default to **audit-only**. If the user wants remediation, propose changes and get explicit approval before running any:\n   - `openclaw security audit --fix`\n   - config edits (`openclaw config set`, Control UI Config, `config.apply` \u002F `config.patch`)\n   - firewall edits\n   - restarts\n\n## What “good” looks like (target posture)\n\n- Gateway binds to loopback unless there is a strong reason otherwise.\n- Strong Gateway auth (token\u002Fpassword, rotated) is enabled.\n- No public internet exposure (no open security group \u002F port-forward \u002F Tailscale Funnel \u002F accidental reverse proxy bypass).\n- DMs require pairing or strict allowlists; group chats require mention-gating.\n- Tooling is least-privilege (deny\u002Fask for exec; workspace-only FS; deny control-plane tools on untrusted surfaces).\n- Secrets are stored with tight filesystem permissions; logs\u002Ftranscripts have a retention plan.\n- Plugins\u002Fskills are explicitly trusted, minimal, and protected from unauthorised modification.\n\n## Workflow\n\n### Step 0 — Establish context (fast)\n\nCollect just enough context to pick the right audit path:\n\n- Where is OpenClaw running?\n  - **Mac mini** (always-on home host)\n  - **Personal laptop** (macOS\u002FWindows\u002FLinux)\n  - **Docker** (local or on a VM)\n  - **Cloud host** (AWS EC2 \u002F VPS)\n- Install type?\n  - native install vs Docker Compose vs source checkout\n- Do we have **local shell access**?\n  - **Claude Code**: can run commands directly\n  - **Claude.ai**: user must run commands and paste outputs\n\nThen pick one of these modes:\n\n- **Mode A: Assisted self-audit (Claude.ai)** — user runs commands\u002Fscripts, shares redacted output.\n- **Mode B: Automated local audit (Claude Code)** — run scripts directly and generate a report.\n\n---\n\n## Mode A — Assisted self-audit (no local shell access)\n\nAsk the user to run the following on the OpenClaw host and paste results.\n\n> If they are worried about sharing output, reassure them that `openclaw status --all` and `openclaw security audit --json` are intended to be shareable. If they still refuse, provide high-level hardening guidance without requesting outputs.\n\n### 1) Collect OpenClaw posture (recommended minimum)\n\n```bash\nopenclaw --version\nopenclaw status --all\nopenclaw doctor\nopenclaw security audit --json\nopenclaw security audit --deep --json\n```\n\nHelpful extras (for supply-chain + snapshot context):\n\n```bash\nopenclaw health --json\nopenclaw skills list --eligible --json\nopenclaw plugins list --json\n```\n\nIf the user is comfortable sharing a config for review, prefer **targeted** reads over full-file sharing:\n\n```bash\nopenclaw config get gateway.bind\nopenclaw config get gateway.auth.mode\nopenclaw config get discovery.mdns.mode\nopenclaw config get session.dmScope\n```\n\nIf they *must* share the full config, remind them OpenClaw config is **JSON5** (`~\u002F.openclaw\u002Fopenclaw.json`), not strict JSON, and have them use the redaction script:\n\n```bash\npython3 scripts\u002Fredact_openclaw_config.py ~\u002F.openclaw\u002Fopenclaw.json > openclaw.json5.redacted\n```\n\n### 2) Collect host\u002Fnetwork posture (pick the right block)\n\n**macOS (Mac mini \u002F Mac laptop):**\n```bash\nwhoami\nsw_vers\nuname -a\nlsof -nP -iTCP -sTCP:LISTEN\n\u002Fusr\u002Flibexec\u002FApplicationFirewall\u002Fsocketfilterfw --getglobalstate\n\u002Fusr\u002Flibexec\u002FApplicationFirewall\u002Fsocketfilterfw --getstealthmode\nfdesetup status || true\n```\n\n**Linux (EC2\u002FVPS):**\n```bash\nwhoami\ncat \u002Fetc\u002Fos-release\nuname -a\nss -ltnp\nsudo ufw status verbose || true\nsudo nft list ruleset || true\nsudo iptables -S || true\n```\n\n**Docker deployments (local or on a VM):**\n```bash\ndocker ps\n# confirm published listeners (should be 127.0.0.1:18789, not 0.0.0.0:18789)\ndocker port openclaw-gateway 18789 || true\ndocker compose ps || true\n```\n\n---\n\n## Mode B — Automated local audit (Claude Code)\n\nIf shell tools are available, run:\n\n```bash\nbash scripts\u002Fcollect_openclaw_audit.sh --out .\u002Fopenclaw-audit\npython3 scripts\u002Frender_report.py --input .\u002Fopenclaw-audit --output .\u002Fopenclaw-security-report.md\n```\n\nThen open `openclaw-security-report.md`, refine wording, and present the report to the user.\n\n---\n\n## Analysis and reporting\n\n### Step 1 — Triage findings by severity\n\nUse OpenClaw’s audit output as the source of truth. Prioritise in this order:\n\n1. **Anything “open” + tools enabled** (lock down DMs\u002Fgroups first; then tighten tools\u002Fsandbox)\n2. **Public network exposure** (LAN bind, Tailscale Funnel, reverse proxy mistakes, missing auth)\n3. **Browser control \u002F nodes exposure** (treat like operator access)\n4. **File permissions** on `~\u002F.openclaw` (config\u002Fstate\u002Fcredentials)\n5. **Plugins\u002Fextensions + dynamic skill loading** (supply chain)\n6. **Model hygiene + prompt-injection resilience** (important, but secondary to access control)\n\n### Step 2 — Translate findings into a clear risk narrative\n\nFor each finding, capture:\n\n- **What was observed** (quote `checkId` + summary; do not paste secrets)\n- **Why it matters**\n- **Impact** (data exfiltration, account takeover, unauthorised actions, persistence)\n- **Concrete remediation steps**\n- **Verification** (usually rerun `openclaw security audit --deep`)\n\n### Step 3 — Recommend a secure baseline configuration (when appropriate)\n\nOffer the baseline in `references\u002Fopenclaw-baseline-config.md` as a starting point and adapt based on needs:\n\n- single-user vs shared inbox\n- required tools\n- remote access method (SSH tunnel \u002F Tailscale Serve)\n\n### Step 4 — Platform-specific guidance\n\nUse the platform playbooks:\n\n- `references\u002Fplatform-mac-mini.md`\n- `references\u002Fplatform-personal-laptop.md`\n- `references\u002Fplatform-aws-ec2.md`\n- `references\u002Fplatform-docker.md`\n\n---\n\n## Deliverable\n\nProduce a Markdown security audit report using `assets\u002Freport-template.md`:\n\n- Executive summary (1–2 paragraphs)\n- Environment overview (host type, install type, exposure)\n- Findings table (Critical\u002FHigh\u002FMedium\u002FLow)\n- Remediation plan (sequenced)\n- Verification steps\n- Residual risk + operational practices (patching, token rotation, retention)\n\n---\n\n## Common remediation patterns (use with care)\n\n### Network exposure\n\n- Prefer Gateway bound to loopback and remote access via **SSH tunnel** or **Tailscale Serve**.\n- Avoid **Tailscale Funnel** for the Gateway.\n- Never expose the Gateway control UI directly to the public internet.\n- If you must run a reverse proxy, configure **trusted proxies** and block direct access to the Gateway port.\n\n### Identity & access\n\n- Use token\u002Fpassword auth; rotate periodically and after any suspected leakage.\n- If you need a token generated in automation: `openclaw doctor --generate-gateway-token`.\n- Require DM pairing or strict allowlists.\n- In groups, require mention gating; avoid always-on bots in large groups.\n\n### Control-plane tools\n\n- The `gateway` and `cron` tool surfaces can create persistence (config changes, scheduled jobs). Deny them on any surface that handles untrusted content.\n\n### Least privilege tools\n\n- Keep filesystem tools workspace-only.\n- Keep exec denied or “ask always” unless the user is present and approving actions.\n- Treat browser control like operator access; keep it tailnet\u002Flocal only.\n\n### Secrets, transcripts, and logs\n\n- Tight permissions on `~\u002F.openclaw` and `~\u002F.openclaw\u002Fopenclaw.json`.\n- Treat session transcripts as sensitive (`~\u002F.openclaw\u002Fagents\u002F\u003CagentId>\u002Fsessions\u002F*.jsonl`).\n- Gateway logs default under `\u002Ftmp\u002Fopenclaw\u002Fopenclaw-YYYY-MM-DD.log` (or `logging.file`).\n- Keep redaction enabled and set a retention\u002Fpruning policy.\n\n---\n\n## Troubleshooting\n\n### “openclaw: command not found”\n\n- Confirm OpenClaw CLI is installed and on PATH.\n- If using Windows, prefer WSL2.\n- Re-run installation via the official method for your platform; then re-run `openclaw --version`.\n\n### “refusing to bind ... without auth”\n\n- Expected guardrail: non-loopback binds require auth. Do not bypass.\n\n### Audit output is empty or inconsistent\n\n- Ensure the Gateway is running: `openclaw gateway status`\n- Re-run with `--deep` for best-effort live probing.\n\n---\n\n## Trigger tests (for skill author)\n\nShould trigger:\n\n- “Can you audit my OpenClaw setup for security?”\n- “My OpenClaw gateway is on port 18789 — is that safe?”\n- “I deployed OpenClaw on EC2; give me a hardening checklist.”\n- “Run openclaw security audit and interpret the results.”\n- “I’m running OpenClaw in Docker—how do I lock it down?”\n\nShould NOT trigger:\n\n- General macOS security advice unrelated to OpenClaw\n- Generic AWS hardening unrelated to OpenClaw\n- Unrelated software audits\n",{"name":5,"license":14,"metadata":23,"description":7,"compatibility":27},{"author":24,"version":25,"upstream":26},"community","2.0.0","OpenClaw docs + security guidance (updated Feb 2026)","Works best in Claude Code (local shell access). In Claude.ai, instruct the user to run the included scripts\u002Fcommands locally and share redacted outputs (prefer `openclaw status --all` and `openclaw security audit --json`).","main","54afe4b75ed2eb8f0feabb2f5b57dfa4bcbb3c44",".agents\u002Fskills\u002Faudit-openclaw-security","https:\u002F\u002Fgithub.com\u002Fmeghal86\u002Fstacksignal","https:\u002F\u002Fgithub.com\u002Fmeghal86","2026-05-17 14:11:28",{"source":8,"skillName":5,"commands":35,"prompt":45,"zipAvailable":46},[36,39,42],{"tool":37,"command":38},"npx","npx skills add meghal86\u002Fstacksignal --skill audit-openclaw-security",{"tool":40,"command":41},"bunx","bunx skills add meghal86\u002Fstacksignal --skill audit-openclaw-security",{"tool":43,"command":44},"pnpm","pnpm dlx skills add meghal86\u002Fstacksignal --skill audit-openclaw-security","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fmeghal86\u002Fstacksignal\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Faudit-openclaw-security\u002FSKILL.md",true,[48,54,60,64,68,72,76,80,84,88,92,98,102],{"path":49,"resourceType":50,"resourceTypeDesc":51,"sizeBytes":52,"sourceHtmlUrl":53},"_meta.json","REFERENCE","参考文档",309,"https:\u002F\u002Fgithub.com\u002Fmeghal86\u002Fstacksignal\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Faudit-openclaw-security\u002F_meta.json",{"path":55,"resourceType":56,"resourceTypeDesc":57,"sizeBytes":58,"sourceHtmlUrl":59},"assets\u002Fopenclaw_checkid_map.json","ASSET","资产文件",3620,"https:\u002F\u002Fgithub.com\u002Fmeghal86\u002Fstacksignal\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Faudit-openclaw-security\u002Fassets\u002Fopenclaw_checkid_map.json",{"path":61,"resourceType":56,"resourceTypeDesc":57,"sizeBytes":62,"sourceHtmlUrl":63},"assets\u002Freport-template.md",1814,"https:\u002F\u002Fgithub.com\u002Fmeghal86\u002Fstacksignal\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Faudit-openclaw-security\u002Fassets\u002Freport-template.md",{"path":65,"resourceType":50,"resourceTypeDesc":51,"sizeBytes":66,"sourceHtmlUrl":67},"references\u002Fcommand-cheatsheet.md",1169,"https:\u002F\u002Fgithub.com\u002Fmeghal86\u002Fstacksignal\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Faudit-openclaw-security\u002Freferences\u002Fcommand-cheatsheet.md",{"path":69,"resourceType":50,"resourceTypeDesc":51,"sizeBytes":70,"sourceHtmlUrl":71},"references\u002Fopenclaw-audit-checks.md",2909,"https:\u002F\u002Fgithub.com\u002Fmeghal86\u002Fstacksignal\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Faudit-openclaw-security\u002Freferences\u002Fopenclaw-audit-checks.md",{"path":73,"resourceType":50,"resourceTypeDesc":51,"sizeBytes":74,"sourceHtmlUrl":75},"references\u002Fopenclaw-baseline-config.md",2527,"https:\u002F\u002Fgithub.com\u002Fmeghal86\u002Fstacksignal\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Faudit-openclaw-security\u002Freferences\u002Fopenclaw-baseline-config.md",{"path":77,"resourceType":50,"resourceTypeDesc":51,"sizeBytes":78,"sourceHtmlUrl":79},"references\u002Fplatform-aws-ec2.md",1816,"https:\u002F\u002Fgithub.com\u002Fmeghal86\u002Fstacksignal\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Faudit-openclaw-security\u002Freferences\u002Fplatform-aws-ec2.md",{"path":81,"resourceType":50,"resourceTypeDesc":51,"sizeBytes":82,"sourceHtmlUrl":83},"references\u002Fplatform-docker.md",2627,"https:\u002F\u002Fgithub.com\u002Fmeghal86\u002Fstacksignal\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Faudit-openclaw-security\u002Freferences\u002Fplatform-docker.md",{"path":85,"resourceType":50,"resourceTypeDesc":51,"sizeBytes":86,"sourceHtmlUrl":87},"references\u002Fplatform-mac-mini.md",2555,"https:\u002F\u002Fgithub.com\u002Fmeghal86\u002Fstacksignal\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Faudit-openclaw-security\u002Freferences\u002Fplatform-mac-mini.md",{"path":89,"resourceType":50,"resourceTypeDesc":51,"sizeBytes":90,"sourceHtmlUrl":91},"references\u002Fplatform-personal-laptop.md",1406,"https:\u002F\u002Fgithub.com\u002Fmeghal86\u002Fstacksignal\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Faudit-openclaw-security\u002Freferences\u002Fplatform-personal-laptop.md",{"path":93,"resourceType":94,"resourceTypeDesc":95,"sizeBytes":96,"sourceHtmlUrl":97},"scripts\u002Fcollect_openclaw_audit.sh","SCRIPT","脚本",4005,"https:\u002F\u002Fgithub.com\u002Fmeghal86\u002Fstacksignal\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Faudit-openclaw-security\u002Fscripts\u002Fcollect_openclaw_audit.sh",{"path":99,"resourceType":94,"resourceTypeDesc":95,"sizeBytes":100,"sourceHtmlUrl":101},"scripts\u002Fredact_openclaw_config.py",5189,"https:\u002F\u002Fgithub.com\u002Fmeghal86\u002Fstacksignal\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Faudit-openclaw-security\u002Fscripts\u002Fredact_openclaw_config.py",{"path":103,"resourceType":94,"resourceTypeDesc":95,"sizeBytes":104,"sourceHtmlUrl":105},"scripts\u002Frender_report.py",10188,"https:\u002F\u002Fgithub.com\u002Fmeghal86\u002Fstacksignal\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Faudit-openclaw-security\u002Fscripts\u002Frender_report.py",[107,116,124,132,140,147],{"skillNo":108,"name":109,"nameZh":6,"description":110,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":111,"sourceHtmlUrl":112,"categories":113,"keywords":114,"publishedAt":115},"01M78XH0FRP65NYEDHF2BT7NB1","zvukogram","Text-to-Speech via Zvukogram API with SSML support. Use when you need to generate speech from text, create podcasts, voice notifications, or work with audio. Supports speed control, stress marks, English word transcription, and audio fragment merging.","https:\u002F\u002Fgithub.com\u002Fmeghal86\u002Fstacksignal\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fzvukogram\u002FLICENSE.txt","https:\u002F\u002Fgithub.com\u002Fmeghal86\u002Fstacksignal\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fzvukogram\u002FSKILL.md",[],[],"2026-11-15 14:19:17",{"skillNo":117,"name":118,"nameZh":6,"description":119,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":120,"sourceHtmlUrl":121,"categories":122,"keywords":123,"publishedAt":115},"01M78XEND5ASXYB6MN8AQQGJAF","zohomail","Zoho Mail API integration with managed OAuth. Send, receive, and manage emails, folders, and labels.\nUse this skill when users want to send emails, read messages, manage folders, or work with email labels in Zoho Mail.\nFor other third party apps, use the api-gateway skill (https:\u002F\u002Fclawhub.ai\u002Fbyungkyu\u002Fapi-gateway).\nRequires network access and valid Maton API key.","https:\u002F\u002Fgithub.com\u002Fmeghal86\u002Fstacksignal\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fzohomail\u002FLICENSE.txt","https:\u002F\u002Fgithub.com\u002Fmeghal86\u002Fstacksignal\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fzohomail\u002FSKILL.md",[],[],{"skillNo":125,"name":126,"nameZh":6,"description":127,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":128,"sourceHtmlUrl":129,"categories":130,"keywords":131,"publishedAt":115},"01M78XEBBV3C6DRZFAJXP6RDKS","zoho-recruit","Zoho Recruit API integration with managed OAuth. Manage candidates, job openings, interviews, and recruitment workflows.\nUse this skill when users want to read, create, update, or search recruitment data like candidates, job openings, interviews, and applications in Zoho Recruit.\nFor other third party apps, use the api-gateway skill (https:\u002F\u002Fclawhub.ai\u002Fbyungkyu\u002Fapi-gateway).\nRequires network access and valid Maton API key.","https:\u002F\u002Fgithub.com\u002Fmeghal86\u002Fstacksignal\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fzoho-recruit\u002FLICENSE.txt","https:\u002F\u002Fgithub.com\u002Fmeghal86\u002Fstacksignal\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fzoho-recruit\u002FSKILL.md",[],[],{"skillNo":133,"name":134,"nameZh":6,"description":135,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":136,"sourceHtmlUrl":137,"categories":138,"keywords":139,"publishedAt":115},"01M78XE5CQG75DZWYSKGXFR99T","zoho-people","Zoho People API integration with managed OAuth. Manage employees, departments, designations, attendance, and leave.\nUse this skill when users want to read, create, update, or query HR data like employees, departments, designations, and forms in Zoho People.\nFor other third party apps, use the api-gateway skill (https:\u002F\u002Fclawhub.ai\u002Fbyungkyu\u002Fapi-gateway).\nRequires network access and valid Maton API key.","https:\u002F\u002Fgithub.com\u002Fmeghal86\u002Fstacksignal\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fzoho-people\u002FLICENSE.txt","https:\u002F\u002Fgithub.com\u002Fmeghal86\u002Fstacksignal\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fzoho-people\u002FSKILL.md",[],[],{"skillNo":141,"name":142,"nameZh":6,"description":119,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":143,"sourceHtmlUrl":144,"categories":145,"keywords":146,"publishedAt":115},"01M78XDZKMZAJYFJJFT7MR5ZN2","zoho-mail","https:\u002F\u002Fgithub.com\u002Fmeghal86\u002Fstacksignal\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fzoho-mail\u002FLICENSE.txt","https:\u002F\u002Fgithub.com\u002Fmeghal86\u002Fstacksignal\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fzoho-mail\u002FSKILL.md",[],[],{"skillNo":148,"name":149,"nameZh":6,"description":150,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":151,"sourceHtmlUrl":152,"categories":153,"keywords":154,"publishedAt":115},"01M78XDST8JK0WVDA1F3VPNWXD","zoho-inventory","Zoho Inventory API integration with managed OAuth. Manage items, sales orders, invoices, purchase orders, bills, contacts, and shipments.\nUse this skill when users want to read, create, update, or delete inventory items, sales orders, invoices, purchase orders, bills, or other inventory records in Zoho Inventory.\nFor other third party apps, use the api-gateway skill (https:\u002F\u002Fclawhub.ai\u002Fbyungkyu\u002Fapi-gateway).\nRequires network access and valid Maton API key.","https:\u002F\u002Fgithub.com\u002Fmeghal86\u002Fstacksignal\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fzoho-inventory\u002FLICENSE.txt","https:\u002F\u002Fgithub.com\u002Fmeghal86\u002Fstacksignal\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fzoho-inventory\u002FSKILL.md",[],[]]