{"_id":"@aeokit/skills","_rev":"3-0ca9856093fddf9f48670430244d5df5","name":"@aeokit/skills","dist-tags":{"latest":"0.3.0"},"versions":{"0.1.1":{"name":"@aeokit/skills","version":"0.1.1","keywords":["agent-skills","aeo","codex","cursor","github-copilot","gemini-cli","claude-code","aeokit"],"license":"MIT","_id":"@aeokit/skills@0.1.1","maintainers":[{"name":"aeokit","email":"jackwalkerlabs@gmail.com"}],"homepage":"https://github.com/aeokit-dev/aeo-skills#readme","bugs":{"url":"https://github.com/aeokit-dev/aeo-skills/issues"},"bin":{"aeokit-skills":"bin/aeo-skills.js"},"dist":{"shasum":"37bc08a8da2cc3f2feff1ad11105c04ce3fa9353","tarball":"https://registry.npmjs.org/@aeokit/skills/-/skills-0.1.1.tgz","fileCount":21,"integrity":"sha512-GKhFjaQmCmEwIVVBll7uvn/gsO1Rh7T7waIwmyjqkE4cQUhn4otOYn8/cQL2rOo6YQkxeuU75KE/ZGqhY3j/yA==","signatures":[{"sig":"MEUCIATYUc2Jda92ZbzmMDSrk5gYWrjlEopiEO/RlgHs5tiSAiEAoxmkqG7Ubi3aWRcLVaM7xQUVu7xyLchXiX9w3jDp3Uw=","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":214699},"type":"module","engines":{"node":">=22.19"},"scripts":{"test":"node --test","check":"node --check bin/aeo-skills.js && node --check src/cli.js && node --check src/audit-citations.js && npm run validate:skills && npm run audit:citations && node --test","prepublishOnly":"npm run check","audit:citations":"node src/audit-citations.js","validate:skills":"node src/validate.js"},"_npmUser":{"name":"aeokit","email":"jackwalkerlabs@gmail.com"},"repository":{"url":"git+https://github.com/aeokit-dev/aeo-skills.git","type":"git"},"_npmVersion":"11.19.0","description":"Portable Agent Skills for evidence-backed answer-engine optimization workflows.","directories":{},"_nodeVersion":"26.7.0","dependencies":{"yaml":"^2.9.0"},"publishConfig":{"access":"public"},"_hasShrinkwrap":false,"_npmOperationalInternal":{"tmp":"tmp/skills_0.1.1_1788109471569_0.19925209061177918","host":"s3://npm-registry-packages-npm-production"}},"0.2.0":{"name":"@aeokit/skills","version":"0.2.0","keywords":["agent-skills","aeo","codex","cursor","github-copilot","gemini-cli","claude-code","aeokit"],"license":"MIT","_id":"@aeokit/skills@0.2.0","maintainers":[{"name":"aeokit","email":"jackwalkerlabs@gmail.com"}],"homepage":"https://github.com/aeokit-dev/aeokit-skills#readme","bugs":{"url":"https://github.com/aeokit-dev/aeokit-skills/issues"},"bin":{"aeokit-skills":"bin/aeo-skills.js"},"dist":{"shasum":"079561cbf9dbade8315e40bf72fa67a6209ba7a5","tarball":"https://registry.npmjs.org/@aeokit/skills/-/skills-0.2.0.tgz","fileCount":32,"integrity":"sha512-1Oz9wdRPa9aVWKt4xLHTOiGaUUzb2u08zLn7V+Z83BFbQ7a4Pl8FtxyZD4Q+nd/etCuxTEdhB+noWBHuqaIDGw==","signatures":[{"sig":"MEQCIBXX5DWNgYsBo84SsH1O8yh7Kgqhv26oorv7Tcms71UjAiB3bZ6++I//deM6V/24OHWGZNB4h5Bm71ZdCp0HY5q8JA==","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"attestations":{"url":"https://registry.npmjs.org/-/npm/v1/attestations/@aeokit%2fskills@0.2.0","provenance":{"predicateType":"https://slsa.dev/provenance/v1"}},"unpackedSize":246957},"type":"module","engines":{"node":">=22.19"},"gitHead":"f05bb2aa145934b3e9bd8d7f51bfd9282d773ba2","scripts":{"test":"node --test","check":"node --check bin/aeo-skills.js && node --check src/cli.js && node --check src/audit-citations.js && npm run validate:skills && npm run audit:citations && node --test","prepublishOnly":"npm run check","audit:citations":"node src/audit-citations.js","sync:aeokit-api":"node src/sync-api-export.js","validate:skills":"node src/validate.js"},"_npmUser":{"name":"GitHub Actions","email":"npm-oidc-no-reply@github.com","trustedPublisher":{"id":"github","oidcConfigId":"oidc:ba5d5d0d-09f9-42e7-82b4-15cd6249df8e"}},"repository":{"url":"git+https://github.com/aeokit-dev/aeokit-skills.git","type":"git"},"_npmVersion":"12.0.2","description":"Portable Agent Skills for evidence-backed answer-engine optimization workflows.","directories":{},"_nodeVersion":"24.19.0","dependencies":{"yaml":"^2.9.0"},"publishConfig":{"access":"public"},"_hasShrinkwrap":false,"_npmOperationalInternal":{"tmp":"tmp/skills_0.2.0_1788221759524_0.4551690834262092","host":"s3://npm-registry-packages-npm-production"}},"0.3.0":{"name":"@aeokit/skills","version":"0.3.0","description":"Portable Agent Skills for evidence-backed answer-engine optimization workflows.","type":"module","bin":{"aeokit-skills":"bin/aeo-skills.js"},"scripts":{"test":"node --test","validate:skills":"node src/validate.js","sync:aeokit-api":"node src/sync-api-export.js","audit:citations":"node src/audit-citations.js","check":"node --check bin/aeo-skills.js && node --check src/cli.js && node --check src/audit-citations.js && npm run validate:skills && npm run audit:citations && node --test","prepublishOnly":"npm run check"},"engines":{"node":">=22.19"},"keywords":["agent-skills","aeo","codex","cursor","github-copilot","gemini-cli","claude-code","aeokit"],"repository":{"type":"git","url":"git+https://github.com/aeokit-dev/aeokit-skills.git"},"homepage":"https://github.com/aeokit-dev/aeokit-skills#readme","bugs":{"url":"https://github.com/aeokit-dev/aeokit-skills/issues"},"license":"MIT","publishConfig":{"access":"public"},"dependencies":{"yaml":"^2.9.0"},"gitHead":"ad95582dc5a6182960c39b5a27327a7bec68c26b","_id":"@aeokit/skills@0.3.0","_nodeVersion":"24.19.0","_npmVersion":"12.0.2","dist":{"integrity":"sha512-mKbVr3qOMRxDAH0cKm9w8bBrjAaf9sPyJhItVLQqBfyQtZWhIUQ9FpEbd9Rf+9byJ8T5cJLAtI6n7iAJ+tf/7w==","shasum":"38f95103a75c5983a092430da9161df7cd9c1393","tarball":"https://registry.npmjs.org/@aeokit/skills/-/skills-0.3.0.tgz","fileCount":35,"unpackedSize":254424,"attestations":{"url":"https://registry.npmjs.org/-/npm/v1/attestations/@aeokit%2fskills@0.3.0","provenance":{"predicateType":"https://slsa.dev/provenance/v1"}},"signatures":[{"keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U","sig":"MEQCIDyTX8VypInwnJn4g4pqfxGGh1pz5QB/sgF0qDk3fk8sAiAIMrxnGekr/zjsDFUSl+PiiqnAIVOWMSyGrGBwfQRVpg=="}]},"_npmUser":{"name":"GitHub Actions","email":"npm-oidc-no-reply@github.com","trustedPublisher":{"id":"github","oidcConfigId":"oidc:ba5d5d0d-09f9-42e7-82b4-15cd6249df8e"}},"directories":{},"maintainers":[{"name":"aeokit","email":"jackwalkerlabs@gmail.com"}],"_npmOperationalInternal":{"host":"s3://npm-registry-packages-npm-production","tmp":"tmp/skills_0.3.0_1788229542945_0.3362694538246138"},"_hasShrinkwrap":false}},"time":{"created":"2026-08-30T17:04:31.378Z","modified":"2026-09-01T02:25:43.447Z","0.1.1":"2026-08-30T17:04:31.730Z","0.2.0":"2026-09-01T00:15:59.690Z","0.3.0":"2026-09-01T02:25:43.119Z"},"bugs":{"url":"https://github.com/aeokit-dev/aeokit-skills/issues"},"license":"MIT","homepage":"https://github.com/aeokit-dev/aeokit-skills#readme","keywords":["agent-skills","aeo","codex","cursor","github-copilot","gemini-cli","claude-code","aeokit"],"repository":{"type":"git","url":"git+https://github.com/aeokit-dev/aeokit-skills.git"},"description":"Portable Agent Skills for evidence-backed answer-engine optimization workflows.","maintainers":[{"name":"aeokit","email":"jackwalkerlabs@gmail.com"}],"readme":"# AEO Skills\n\nPortable Agent Skills by aeokit for evidence-backed answer-engine optimization. The same `SKILL.md` workflows run in ChatGPT, Codex, Cursor, GitHub Copilot, Gemini CLI, Claude Code, or [AEO Agent](https://github.com/aeokit-dev/aeo-agent).\n\n**Audit the evidence:** [Open the numbered research claim and source ledger](research/report-source.md#citation-and-traceability-registry).\n\n## Included skills\n\n| Skill | Job |\n|---|---|\n| `aeo-audit` | Read-only crawlability, entity, extraction, schema-parity, and evidence audit |\n| `aeo-improve` | Diagnose one buyer question, propose or implement a small patch, and verify it honestly |\n| `aeo-observe` | Run stable buyer prompts and report mentions, citations, uncertainty, and raw evidence |\n\nThe skills use the open Agent Skills folder format. They remain useful without AEO Agent; when aeokit tools are installed, the instructions use them for deterministic checks and reproducible observations.\n\n## Zero to cited\n\nSEO competes to rank a page. AEO competes to make useful evidence available to\nan AI-generated answer. A new, original source can potentially appear in an\nanswer as soon as a live answer engine discovers it and finds it relevant; it\ndoes not necessarily need months of ranking history first. Inclusion is never\nguaranteed, so this repository provides a reproducible workflow rather than a\npromise of instant visibility.\n\nUse the three skills together to follow a source from publication to its first\nobserved citation:\n\n```text\naeo-audit -> aeo-improve -> publish -> aeo-observe\n                                      |\n                                      v\n                 crawler visit -> mention -> citation\n```\n\n- `aeo-audit` checks whether the source is accessible, extractable, and\n  supported by verifiable evidence.\n- `aeo-improve` makes the smallest authorized change for a specific buyer\n  question and verifies the local result.\n- `aeo-observe` repeats stable prompts and records what answer engines actually\n  mention and cite.\n\naeokit is the system of record for auditable answers, citations, mentions, and\ncrawler activity. AEO Agent is the conversational interface for exploring that\nevidence. These skills make the workflow portable across supported agents.\n\n## Install the aeokit plugin\n\nThe plugin is the recommended installation for Codex, ChatGPT, and Claude Code. It keeps the three portable skills together under the `aeokit` identity.\n\nFor Claude Code, add this GitHub repository as a marketplace and install the plugin:\n\n```text\n/plugin marketplace add aeokit-dev/aeokit-skills\n/plugin install aeokit@aeokit\n```\n\nThe skills are then available as `/aeokit:aeo-audit`, `/aeokit:aeo-improve`, and `/aeokit:aeo-observe`.\n\nFor Codex and ChatGPT, install `aeokit` from the universal plugin directory after publication.\n\n## Install standalone skill folders\n\nList the catalog:\n\n```bash\nnpx @aeokit/skills@latest list\n```\n\nInstall one skill into the current repository:\n\n```bash\nnpx @aeokit/skills@latest add aeo-improve --to agents\nnpx @aeokit/skills@latest add aeo-improve --to claude\nnpx @aeokit/skills@latest add aeo-improve --to agent\n```\n\nThe shared `agents` target works with Codex, Cursor, GitHub Copilot, and Gemini CLI. The aliases `codex`, `cursor`, `copilot`, and `gemini` resolve to that same location.\n\nInstall all skills for your user account:\n\n```bash\nnpx @aeokit/skills@latest add --all --to agents --scope user\nnpx @aeokit/skills@latest add --all --to claude --scope user\nnpx @aeokit/skills@latest add --all --to agent --scope user\n```\n\nExisting directories are never replaced unless `--force` is explicit. Use `--dry-run` to inspect destinations first.\n\n## Locations\n\n| Host | Project | User |\n|---|---|---|\n| Codex, Cursor, GitHub Copilot, Gemini CLI | `.agents/skills/<name>/` | `~/.agents/skills/<name>/` |\n| Claude Code | `.claude/skills/<name>/` | `~/.claude/skills/<name>/` |\n| AEO Agent | `.aeokit/skills/<name>/` | `~/.aeokit/skills/<name>/` |\n\nThe shared `.agents/skills/` location is documented by [OpenAI](https://developers.openai.com/codex/build-skills), [Cursor](https://cursor.com/docs/skills), [GitHub Copilot](https://docs.github.com/en/copilot/how-tos/copilot-on-github/customize-copilot/customize-cloud-agent/add-skills), and [Gemini CLI](https://geminicli.com/docs/cli/skills/). Claude Code discovers the same skill format from its own directory. AEO Agent uses that file structure and adds typed AEO tool and evidence contracts.\n\n## Plugin distribution\n\nThe repository root is a skills-only plugin for the OpenAI and Claude plugin systems:\n\n| Ecosystem | Manifest | Skill discovery |\n|---|---|---|\n| ChatGPT and Codex | `.codex-plugin/plugin.json` | `./skills/` declared by the manifest |\n| Claude Code | `.claude-plugin/plugin.json` | `./skills/` discovered from the plugin root |\n\nThe OpenAI plugin exposes the same three skills in ChatGPT and Codex. The Claude Code plugin exposes them with the `aeokit` namespace, such as `/aeokit:aeo-audit`. Neither manifest forks or rewrites the underlying skill instructions.\n\nEach skill also contains the same versioned AeoKit API workflow reference for Claude Code and Codex. Those API-specific files are authored and tested in the main `aeokit` repository, then imported into this repository before release. End users still install only this plugin; no runtime download or second skill installation is required.\n\nThe Claude marketplace catalog is stored at `.claude-plugin/marketplace.json`, so this GitHub repository can be added directly as a marketplace.\n\nThis layout follows the [OpenAI skill and plugin guidance](https://developers.openai.com/codex/build-skills) and [Claude Code plugin guidance](https://code.claude.com/docs/en/plugins). The npm installer remains useful for direct project or user installation without a plugin marketplace.\n\n## Invoke\n\nCodex:\n\n```text\n$aeo-audit Audit this repository for the buyer question \"best open-source AEO testing tools.\"\n```\n\nChatGPT:\n\n```text\nSelect the installed AEO Audit skill with @, then ask it to audit the site for a buyer question.\n```\n\nClaude Code:\n\n```text\n/aeokit:aeo-improve Make this site easier to accurately cite for \"best open-source AEO testing tools.\"\n```\n\nThe namespaced form above is for the plugin. A direct `--to claude` installation uses `/aeo-improve`. Cursor and GitHub Copilot expose installed skills through `/`; Gemini CLI lists and manages them through `/skills`.\n\nAEO Agent loads `aeo-improve` automatically during planning when installed:\n\n```bash\naeo-agent skills\naeo-agent plan --query \"What are the best open-source AEO testing tools?\"\n```\n\n## Design rules\n\n- The optimizing agent does not grade its own work.\n- Every diagnosis is tied to an evidence ID or source URL.\n- Local checks, grounded simulations, and production observations are labeled separately.\n- Skills never invent claims or imply that schema alone creates visibility.\n- Mutating work requires user authorization; deployment is a separate action.\n\n## Research basis\n\nThe skills are maintained against a [source-backed AEO evidence review](research/report-source.md), last researched on 2026-08-29. Every substantive instructional line links to a stable research claim (`C##`), source (`S##`), or repository policy (`P##`) in its [numbered citation ledger](research/report-source.md#citation-and-traceability-registry).\n\nIts core conclusion is that durable AEO combines platform-specific retrieval access, people-first and verifiable content, consistent entity data, and reproducible per-surface measurement. It does not assume that `llms.txt`, special schema, artificial content chunking, or a single prompt observation improves visibility.\n\n## Development\n\nRequires Node.js 22.19 or newer.\n\n```bash\nnpm install\nnpm run sync:aeokit-api -- /path/to/aeokit/agent-skills/api-export\nnpm run audit:citations\nnpm run check\n```\n\nThe sync command writes `skills/.aeokit-api.lock.json` with the API version and SHA-256 digest of every imported file. `npm run check` verifies that lock, so locally edited or stale imported guidance cannot be published accidentally.\n\nValidate the Claude plugin and marketplace with Claude Code 2.1.233 or newer:\n\n```bash\nclaude plugin validate .\n```\n\nValidate the Codex plugin before submission. Public Codex and ChatGPT publication uses the OpenAI Platform plugin submission portal; public Claude publication uses the plugin submission form in Claude.ai or the Anthropic Console.\n\nReleases are automated with Release Please and npm trusted publishing. Conventional `feat:`, `fix:`, and breaking-change commits update a release PR; merging that PR creates the GitHub release and publishes the exact tagged commit to npm with provenance.\n\nOne-time activation requires an organization owner to allow GitHub Actions to create pull requests, then an npm package owner must add a trusted publisher for repository `aeokit-dev/aeokit-skills`, workflow `release.yml`, and environment `npm`. Finally set the repository variable `RELEASE_AUTOMATION_ENABLED` to `true`. The release job remains safely disabled while that variable is absent or false.\n\n## License\n\nMIT\n","readmeFilename":"README.md"}