[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M75PM13APMN16MDZTVK3B6QF":3,"skill-same:01M75PM13APMN16MDZTVK3B6QF":58},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":38,"commitSha":39,"skillDirectory":40,"forksCount":14,"repositoryHtmlUrl":41,"repositoryHomepageUrl":6,"ownerHtmlUrl":42,"ownerBlogUrl":43,"ownerDescription":6,"updatedAt":44,"install":45,"resources":57},"01M75PM13APMN16MDZTVK3B6QF","llm-setup-audit",null,"Use when reviewing or hardening Claude Code agent configuration (.claude\u002Fsettings.json \u002F settings.local.json) — before committing config changes or onboarding a repo. Covers permissions, sandbox, hooks, MCP servers, plugins, committed secrets, and secrets-file gitignore. For application\u002Finfra security use web-security-audit.","Aliendreamer\u002Fai.skills","Aliendreamer","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F26008043?v=4","01M75PJZKADCYNHHZPMW7363F6","Teodor Stefanov","01M75PJZKYGDX0TDM0X5TX24G0",0,"MIT","https:\u002F\u002Fgithub.com\u002FAliendreamer\u002Fai.skills\u002Fblob\u002Fmain\u002Fskills\u002Fagent-setup\u002Fllm-setup-audit\u002FSKILL.md",[],[],"2026-11-12 13:55:43",false,"---\nname: llm-setup-audit\ndescription:\n  \"Use when reviewing or hardening Claude Code agent configuration (.claude\u002Fsettings.json \u002F settings.local.json) —\n  before committing config changes or onboarding a repo. Covers permissions, sandbox, hooks, MCP servers, plugins,\n  committed secrets, and secrets-file gitignore. For application\u002Finfra security use web-security-audit.\"\ntype: skill\ndisable-model-invocation: false\nuser-invocable: true\ntags: [security, audit, claude-code, configuration]\nagents: [claude, codex, cursor, gemini, copilot]\nversion: 0.1.0\n---\n\n# Agent Setup Audit (.claude\u002Fsettings.json)\n\n## Overview\n\nChecks the **Claude Code harness configuration** for permission-bypass and supply-chain risk: who can run what, what\nruns automatically (hooks), what code is pulled in (MCP servers, plugins), and whether secrets leaked into a committed\nfile. Run each from the repo root, report **PASS** or **FLAG**, investigate every FLAG.\n\nThis audits **how the agent is allowed to act**. For the security of the code\u002Finfra you ship, use\n**web-security-audit**.\n\n## When to use\n\n- Before committing changes to `.claude\u002Fsettings.json` \u002F `settings.local.json`.\n- When onboarding or reviewing a repo whose agent config you didn't write.\n- When an automated commit-review flags a permission\u002Fbypass issue.\n\n## Checks\n\n### 1. No blanket Bash allow\n\n```bash\njq -e '.permissions.allow[]? | select(. == \"Bash\" or . == \"Bash(*)\")' .claude\u002Fsettings.json\n```\n\nPASS: exit ≠ 0 (no match). Bash should be a **narrow toolchain allowlist** (`Bash(git *)`, `Bash(dotnet *)`, …). FLAG:\n`Bash(*)` \u002F `Bash` — any-command bypass.\n\n### 2. Permission mode isn't silently permissive\n\n```bash\njq -r '.permissions.defaultMode \u002F\u002F \"default\"' .claude\u002Fsettings.json\n```\n\nFLAG (unless deliberate + documented): `bypassPermissions` or `dontAsk` \u002F `acceptEdits` as the standing default.\n\n### 3. Deny list is present (but not relied upon)\n\n```bash\njq -e '.permissions.deny' .claude\u002Fsettings.json     # exists = ok\n```\n\nA literal-prefix `deny` list (`Bash(rm -rf*)`) is **best-effort only** — bypassed by spacing, chaining (`&&`), or\naliasing. It's defense-in-depth, never the primary control. FLAG: a deny list used _instead of_ an allowlist.\n\n### 4. Sandbox state is intentional\n\n```bash\njq -r '.sandbox.enabled \u002F\u002F \"unset\"' .claude\u002Fsettings.json\n```\n\nThe value must be deliberate and documented (commit\u002FCLAUDE.md\u002Fmemory). If `false`, the Bash allowlist (check 1) is the\nguardrail. If `true`, confirm it actually engages on this host (it can silently fall back to unsandboxed without\n`bwrap`\u002F`socat`).\n\n### 5. No secrets in the committed settings\n\n```bash\njq -e '(.env \u002F\u002F {}) | to_entries[] | select(.value|test(\"(?i)(secret|token|api[_-]?key|password)\"))' .claude\u002Fsettings.json\ngit check-ignore .claude\u002Fsettings.json    # PASS: NOT ignored → it's committed → must hold no secrets\n```\n\n`settings.json` is git-tracked. FLAG: secrets in `.env`, in a hook `command`, or in `mcpServers` args. Real secrets\nbelong in **`.claude\u002Fsettings.local.json`** (which must be gitignored).\n\nFor agent behavioral rules around reading and echoing secrets at runtime, see **`secrets-safety`**.\n\n### 6. Hooks are safe (they run automatically = RCE surface)\n\n```bash\njq -r '(.hooks \u002F\u002F {}) | to_entries[] | .key as $e | .value[].hooks[] | \"\\($e): \\(.command \u002F\u002F .url \u002F\u002F .type)\"' .claude\u002Fsettings.json\n```\n\nReview each: does it write **outside** the repo, `curl`\u002Fexfiltrate, or pipe untrusted tool input into a shell? A hook\nfires on every matching event — a hostile one is silent code execution. FLAG: network calls to unknown hosts, writes to\n`~`\u002Fsystem paths, or `eval` of tool output.\n\n### 7. MCP servers are from trusted sources\n\n```bash\njq -r '(.mcpServers \u002F\u002F {}) | to_entries[] | \"\\(.key): \\(.value.command) \\(.value.args|join(\" \"))\"' .claude\u002Fsettings.json\njq -r '.enableAllProjectMcpServers \u002F\u002F false' .claude\u002Fsettings.json\n```\n\nFLAG: a server pulling from an unpinned\u002Funtrusted source, or `enableAllProjectMcpServers: true` (auto-approves every\nserver in `.mcp.json`) without having reviewed `.mcp.json`.\n\n### 8. Plugins \u002F marketplaces are trusted\n\n```bash\njq -r '(.enabledPlugins \u002F\u002F {}) | keys[]' .claude\u002Fsettings.json\njq -r '(.extraKnownMarketplaces \u002F\u002F {}) | to_entries[] | \"\\(.key): \\(.value.source)\"' .claude\u002Fsettings.json\n```\n\nFLAG: a plugin from a marketplace you don't recognize\u002Fcontrol.\n\n### 9. The file is valid JSON\n\n```bash\njq -e . .claude\u002Fsettings.json >\u002Fdev\u002Fnull && echo PASS\n[ -f .claude\u002Fsettings.local.json ] && jq -e . .claude\u002Fsettings.local.json >\u002Fdev\u002Fnull\n```\n\n**Invalid JSON silently disables every setting in that file** — including your permission and hook config. Always\nvalidate after editing.\n\n### 10. Secrets source is gitignored\n\nThe user's secrets may live as a single file, a folder, or a file pattern. Check whichever\napplies (ask the user if unsure):\n\n```bash\n# Single file (default: config.conf)\ngit check-ignore -v config.conf\n\n# Folder\ngit check-ignore -v secrets\u002F\n\n# File pattern (e.g. .env.*)\ngit ls-files --others --exclude-standard | grep -E '^\\.env\\.' | head -5\n# All matches should be untracked (gitignored); any tracked file is a FLAG\n```\n\nPASS: the secrets source is fully gitignored. FLAG: any file, folder, or pattern match that\nis **not** gitignored — it can be committed and leaked. The sandbox filesystem restriction is\na runtime guardrail, not a substitute for keeping secrets out of version control.\n\n## Report format\n\nList checks `1–10` as **PASS** \u002F **FLAG** with the offending key for any FLAG and a one-line fix.\n\n## Common mistakes\n\n| Mistake                                       | Why it's wrong                                                      |\n| --------------------------------------------- | ------------------------------------------------------------------- |\n| `Bash(*)` \"for a smooth dev loop\"             | Any-command bypass — narrow to a toolchain allowlist                |\n| Trusting the `deny` list to block `rm -rf`    | Prefix denies are bypassable; only an allowlist actually constrains |\n| Secrets in `.claude\u002Fsettings.json`            | It's committed — put secrets in gitignored `settings.local.json`    |\n| A broad hook with no `matcher`\u002Fpath guard     | Fires on everything; perf cost + a standing execution surface       |\n| `enableAllProjectMcpServers: true` unreviewed | Auto-trusts arbitrary servers declared in `.mcp.json`               |\n| Secrets source (file\u002Ffolder\u002F`.env.*`) not gitignored | Sandbox blocks runtime reads; git doesn't — can be committed   |\n","# Agent Setup Audit (.claude\u002Fsettings.json)\n\n## Overview\n\nChecks the **Claude Code harness configuration** for permission-bypass and supply-chain risk: who can run what, what\nruns automatically (hooks), what code is pulled in (MCP servers, plugins), and whether secrets leaked into a committed\nfile. Run each from the repo root, report **PASS** or **FLAG**, investigate every FLAG.\n\nThis audits **how the agent is allowed to act**. For the security of the code\u002Finfra you ship, use\n**web-security-audit**.\n\n## When to use\n\n- Before committing changes to `.claude\u002Fsettings.json` \u002F `settings.local.json`.\n- When onboarding or reviewing a repo whose agent config you didn't write.\n- When an automated commit-review flags a permission\u002Fbypass issue.\n\n## Checks\n\n### 1. No blanket Bash allow\n\n```bash\njq -e '.permissions.allow[]? | select(. == \"Bash\" or . == \"Bash(*)\")' .claude\u002Fsettings.json\n```\n\nPASS: exit ≠ 0 (no match). Bash should be a **narrow toolchain allowlist** (`Bash(git *)`, `Bash(dotnet *)`, …). FLAG:\n`Bash(*)` \u002F `Bash` — any-command bypass.\n\n### 2. Permission mode isn't silently permissive\n\n```bash\njq -r '.permissions.defaultMode \u002F\u002F \"default\"' .claude\u002Fsettings.json\n```\n\nFLAG (unless deliberate + documented): `bypassPermissions` or `dontAsk` \u002F `acceptEdits` as the standing default.\n\n### 3. Deny list is present (but not relied upon)\n\n```bash\njq -e '.permissions.deny' .claude\u002Fsettings.json     # exists = ok\n```\n\nA literal-prefix `deny` list (`Bash(rm -rf*)`) is **best-effort only** — bypassed by spacing, chaining (`&&`), or\naliasing. It's defense-in-depth, never the primary control. FLAG: a deny list used _instead of_ an allowlist.\n\n### 4. Sandbox state is intentional\n\n```bash\njq -r '.sandbox.enabled \u002F\u002F \"unset\"' .claude\u002Fsettings.json\n```\n\nThe value must be deliberate and documented (commit\u002FCLAUDE.md\u002Fmemory). If `false`, the Bash allowlist (check 1) is the\nguardrail. If `true`, confirm it actually engages on this host (it can silently fall back to unsandboxed without\n`bwrap`\u002F`socat`).\n\n### 5. No secrets in the committed settings\n\n```bash\njq -e '(.env \u002F\u002F {}) | to_entries[] | select(.value|test(\"(?i)(secret|token|api[_-]?key|password)\"))' .claude\u002Fsettings.json\ngit check-ignore .claude\u002Fsettings.json    # PASS: NOT ignored → it's committed → must hold no secrets\n```\n\n`settings.json` is git-tracked. FLAG: secrets in `.env`, in a hook `command`, or in `mcpServers` args. Real secrets\nbelong in **`.claude\u002Fsettings.local.json`** (which must be gitignored).\n\nFor agent behavioral rules around reading and echoing secrets at runtime, see **`secrets-safety`**.\n\n### 6. Hooks are safe (they run automatically = RCE surface)\n\n```bash\njq -r '(.hooks \u002F\u002F {}) | to_entries[] | .key as $e | .value[].hooks[] | \"\\($e): \\(.command \u002F\u002F .url \u002F\u002F .type)\"' .claude\u002Fsettings.json\n```\n\nReview each: does it write **outside** the repo, `curl`\u002Fexfiltrate, or pipe untrusted tool input into a shell? A hook\nfires on every matching event — a hostile one is silent code execution. FLAG: network calls to unknown hosts, writes to\n`~`\u002Fsystem paths, or `eval` of tool output.\n\n### 7. MCP servers are from trusted sources\n\n```bash\njq -r '(.mcpServers \u002F\u002F {}) | to_entries[] | \"\\(.key): \\(.value.command) \\(.value.args|join(\" \"))\"' .claude\u002Fsettings.json\njq -r '.enableAllProjectMcpServers \u002F\u002F false' .claude\u002Fsettings.json\n```\n\nFLAG: a server pulling from an unpinned\u002Funtrusted source, or `enableAllProjectMcpServers: true` (auto-approves every\nserver in `.mcp.json`) without having reviewed `.mcp.json`.\n\n### 8. Plugins \u002F marketplaces are trusted\n\n```bash\njq -r '(.enabledPlugins \u002F\u002F {}) | keys[]' .claude\u002Fsettings.json\njq -r '(.extraKnownMarketplaces \u002F\u002F {}) | to_entries[] | \"\\(.key): \\(.value.source)\"' .claude\u002Fsettings.json\n```\n\nFLAG: a plugin from a marketplace you don't recognize\u002Fcontrol.\n\n### 9. The file is valid JSON\n\n```bash\njq -e . .claude\u002Fsettings.json >\u002Fdev\u002Fnull && echo PASS\n[ -f .claude\u002Fsettings.local.json ] && jq -e . .claude\u002Fsettings.local.json >\u002Fdev\u002Fnull\n```\n\n**Invalid JSON silently disables every setting in that file** — including your permission and hook config. Always\nvalidate after editing.\n\n### 10. Secrets source is gitignored\n\nThe user's secrets may live as a single file, a folder, or a file pattern. Check whichever\napplies (ask the user if unsure):\n\n```bash\n# Single file (default: config.conf)\ngit check-ignore -v config.conf\n\n# Folder\ngit check-ignore -v secrets\u002F\n\n# File pattern (e.g. .env.*)\ngit ls-files --others --exclude-standard | grep -E '^\\.env\\.' | head -5\n# All matches should be untracked (gitignored); any tracked file is a FLAG\n```\n\nPASS: the secrets source is fully gitignored. FLAG: any file, folder, or pattern match that\nis **not** gitignored — it can be committed and leaked. The sandbox filesystem restriction is\na runtime guardrail, not a substitute for keeping secrets out of version control.\n\n## Report format\n\nList checks `1–10` as **PASS** \u002F **FLAG** with the offending key for any FLAG and a one-line fix.\n\n## Common mistakes\n\n| Mistake                                       | Why it's wrong                                                      |\n| --------------------------------------------- | ------------------------------------------------------------------- |\n| `Bash(*)` \"for a smooth dev loop\"             | Any-command bypass — narrow to a toolchain allowlist                |\n| Trusting the `deny` list to block `rm -rf`    | Prefix denies are bypassable; only an allowlist actually constrains |\n| Secrets in `.claude\u002Fsettings.json`            | It's committed — put secrets in gitignored `settings.local.json`    |\n| A broad hook with no `matcher`\u002Fpath guard     | Fires on everything; perf cost + a standing execution surface       |\n| `enableAllProjectMcpServers: true` unreviewed | Auto-trusts arbitrary servers declared in `.mcp.json`               |\n| Secrets source (file\u002Ffolder\u002F`.env.*`) not gitignored | Sandbox blocks runtime reads; git doesn't — can be committed   |\n",{"name":5,"tags":24,"type":29,"agents":30,"version":36,"description":7,"user-invocable":37,"disable-model-invocation":20},[25,26,27,28],"security","audit","claude-code","configuration","skill",[31,32,33,34,35],"claude","codex","cursor","gemini","copilot","0.1.0",true,"main","4cd6c294735ac014fa15da54a397d33d6cdbed57","skills\u002Fagent-setup\u002Fllm-setup-audit","https:\u002F\u002Fgithub.com\u002FAliendreamer\u002Fai.skills","https:\u002F\u002Fgithub.com\u002FAliendreamer","www.linkedin.com\u002Fin\u002Fteodor-stefanov-205aba174","2026-11-12 22:47:39",{"source":8,"skillName":5,"commands":46,"prompt":56,"zipAvailable":37},[47,50,53],{"tool":48,"command":49},"npx","npx skills add Aliendreamer\u002Fai.skills --skill llm-setup-audit",{"tool":51,"command":52},"bunx","bunx skills add Aliendreamer\u002Fai.skills --skill llm-setup-audit",{"tool":54,"command":55},"pnpm","pnpm dlx skills add Aliendreamer\u002Fai.skills --skill llm-setup-audit","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FAliendreamer\u002Fai.skills\u002Fblob\u002Fmain\u002Fskills\u002Fagent-setup\u002Fllm-setup-audit\u002FSKILL.md",[],[59,66,73,80,87,94],{"skillNo":60,"name":61,"nameZh":6,"description":62,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":63,"categories":64,"keywords":65,"publishedAt":19},"01M75PQ9W6ZHVTADCMX9JRVEQ8","setup-git-credentials","Use when the user wants Git to stop using a PAT in the remote URL or a plain-text credential store, wants Git Credential Manager (GCM) with OAuth for GitHub, GitLab or Azure DevOps on Ubuntu\u002FDebian or WSL2, or sees 'cannot persist Microsoft authentication token cache securely' \u002F 'using plain-text fallback token cache'. Asks whether this is WSL, which Git hosts and which parts to set up, checks what is already done, sets git config itself after confirmation and hands every sudo or interactive step (apt, gpg, pass init, secret-tool, first sign-in) to the user. Bundles gcm-keyring-session.sh to start or reuse a D-Bus + GNOME Keyring session without systemd. Trigger terms - git credentials, PAT in URL, token in remote, git credential manager, gcm, github auth, gitlab auth, azure devops auth, pass, gpg store, secret service, gnome keyring, msal token cache, wsl git auth.","https:\u002F\u002Fgithub.com\u002FAliendreamer\u002Fai.skills\u002Fblob\u002Fmain\u002Fskills\u002Fworkflow\u002Fsetup-git-credentials\u002FSKILL.md",[],[],{"skillNo":67,"name":68,"nameZh":6,"description":69,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":70,"categories":71,"keywords":72,"publishedAt":19},"01M75PQ16N14SH0T44Q8GDXS01","setup-flow","Use when onboarding a repository or agent to the standard development workflow — installs the required skills from the ai.skills store, installs or updates a MANDATORY workflow block in the agent's instruction file (CLAUDE.md, AGENTS.md, GEMINI.md, Copilot or Cursor rules), AND configures .claude\u002Fsettings.json with Serena MCP, sandbox, permissions, hooks, and plugins. Hooks are wired to the context-hooks kit, which speaks once per session instead of injecting text on every turn; re-running replaces earlier inline hooks rather than adding to them. Also writes .vscode\u002Fsettings.json to hide the empty shell dotfiles an agent sandbox leaves in the repo root. Use when setting up a new repo, switching agents, or refreshing the workflow rules.","https:\u002F\u002Fgithub.com\u002FAliendreamer\u002Fai.skills\u002Fblob\u002Fmain\u002Fskills\u002Fworkflow\u002Fsetup-flow\u002FSKILL.md",[],[],{"skillNo":74,"name":75,"nameZh":6,"description":76,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":77,"categories":78,"keywords":79,"publishedAt":19},"01M75PPVEK6WWJEBA9PYYR06R6","developer-flow","Use when starting, implementing, or completing any feature, change, or fix in any software project — at feature start, before writing code, and before claiming a change is done, fixed, working, or green. A strict ordered flow with a gate per phase; stack-agnostic and repo-agnostic (OpenSpec optional, gates discovered).","https:\u002F\u002Fgithub.com\u002FAliendreamer\u002Fai.skills\u002Fblob\u002Fmain\u002Fskills\u002Fworkflow\u002Fdeveloper-flow\u002FSKILL.md",[],[],{"skillNo":81,"name":82,"nameZh":6,"description":83,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":84,"categories":85,"keywords":86,"publishedAt":19},"01M75PPPGQP9MV3BAHAP2W2KVV","daily-log-check","Use to check whether everyone on duty filed their Azure DevOps time-log Tasks for a given day, per platform — reports tickets, hours and activities per person against the headcount you give. Fires on \"did everyone log yesterday\", \"check web 4 mobile 3\", \"who is missing for 8.09\", \"daily log check\". Read-only. Org and project come from configuration.","https:\u002F\u002Fgithub.com\u002FAliendreamer\u002Fai.skills\u002Fblob\u002Fmain\u002Fskills\u002Fworkflow\u002Fdaily-log-check\u002FSKILL.md",[],[],{"skillNo":88,"name":89,"nameZh":6,"description":90,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":91,"categories":92,"keywords":93,"publishedAt":19},"01M75PPJFQSRGSRPF8RR0Z8CRS","daily-activity-log","Use at end of day to file the Azure DevOps time-log Tasks — one closed Task per Activity worked that day, Original Estimate in decimal hours, parented to the current sprint's Feature. Fires on \"log my day\", \"fill my tasks\", \"close my daily tasks\", \"I had 2h meetings and 4h development\". Org and project come from configuration.","https:\u002F\u002Fgithub.com\u002FAliendreamer\u002Fai.skills\u002Fblob\u002Fmain\u002Fskills\u002Fworkflow\u002Fdaily-activity-log\u002FSKILL.md",[],[],{"skillNo":95,"name":96,"nameZh":6,"description":97,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":98,"categories":99,"keywords":100,"publishedAt":19},"01M75PPFK8HSW0VB92VRMWAAJE","conventional-commits","Write commit messages that follow the Conventional Commits specification so tooling can derive versions and changelogs.","https:\u002F\u002Fgithub.com\u002FAliendreamer\u002Fai.skills\u002Fblob\u002Fmain\u002Fskills\u002Fworkflow\u002Fconventional-commits\u002FSKILL.md",[],[]]