agentleFS
oegeyilmaz9oegeyilmaz9/seo-aeo-geo-ultimate/skills/seo-agentic/SKILL.md

seo-agentic

Skill6 starsChanged 3 months agoApache-2.0View on GitHub

Use when auditing whether an AI agent can safely understand and complete a web task through the rendered interface, DOM, accessibility tree, forms, authentication, consent, state changes, and documented protocol capabilities such as UCP, ACP, MCP, or A2A; treat this as task readiness and accessibility, not a ranking factor or guaranteed agent usage.

4 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

5 rules. 0 are common, 4 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 4 edited versions we read keep these.

  • 4

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Read markdown-delivery-contract.md for every task.

    line 10
  2. Distinctive:

    Read agent-task-protocol.md before a formal audit.

    line 17
  3. Distinctive:

    Verify semantic controls, stable identifiers/state, keyboard operation, loading/error recovery, consent, authentication, rate limits, and …

    line 23
  4. Distinctive:

    Test task completion with bounded safe inputs in an authorized environment.

    line 25
  5. Distinctive:

    Do not call semantic HTML a ranking factor, protocol support a visibility requirement, or successful automation evidence of universal agent …

    line 32

“Common” is measured: kept by at least 6 in 10 of the 4 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Agentic Web Readiness

Markdown delivery contract

Read markdown-delivery-contract.md for every task. Write a substantive .md document for every material work step and register its direct destination in the work index.md. For public-site work, apply the owner-required Markdown discovery profile across the full eligible inventory; preserve native standards and immutable evidence formats.

Lane outputs: agent-journey.md; task-verification.md (combine small related steps with stable section links). Test read-only knowledge tasks starting from llms.txt and their direct Markdown destinations. Keep UI state, forms, consent and transaction authorization in their existing surfaces; a knowledge export does not authorize actions.

Audit whether a permitted agent can perceive, understand, and safely complete a bounded task using the same truthful interface and controls available to users.

Read agent-task-protocol.md before a formal audit.

Procedure

  1. Define one user-authorized task, start/end states, account/access assumptions, locale/device, risk class, owner, and success evidence.
  2. Capture screenshot, DOM, accessibility tree, accessible names/roles/states, focus order, form labels/errors, dynamic state, network-independent confirmations, and any approved agent trace.
  3. Verify semantic controls, stable identifiers/state, keyboard operation, loading/error recovery, consent, authentication, rate limits, and human-readable confirmation.
  4. Require explicit confirmation for purchase, publication, deletion, permission, financial, legal, or other consequential actions. Never weaken security, privacy, bot controls, or anti-abuse systems for convenience.
  5. Test task completion with bounded safe inputs in an authorized environment. Record partial completion, blocked access, ambiguity, and irreversible steps; do not simulate success.
  6. Evaluate UCP, ACP, MCP, A2A, or similar capabilities only against current specifications and a documented consumer. Record discovery location, operations, authentication, version, fallback, owner, and policy boundary.
  7. Route crawl/index concerns to seo-technical, commerce data to seo-commerce, accessibility remediation to the implementation owner, and search/AI measurement to seo-performance or ai-visibility-monitor.
  8. Package findings as SEO Findings 1.1.0 with task_flow or protocol_capability targets and agentic, accessibility, policy, technical, or commerce categories.

Guardrails

Do not call semantic HTML a ranking factor, protocol support a visibility requirement, or successful automation evidence of universal agent compatibility. Do not bypass authentication, CAPTCHA, consent, rate limits, paywalls, robots, or destructive confirmations.

Formal evidence handoff

Package only supported observations as SEO Findings 1.1.0. Let <suite-root> mean ${CLAUDE_PLUGIN_ROOT} in Claude Code. In Codex, read .seo-suite-runtime.json beside this SKILL.md when present and use its suite_root value; otherwise use the absolute repository checkout, then run python "<suite-root>/scripts/validate_seo_findings.py" validate-findings <bundle>/seo-findings.json --bundle <bundle>. Keep the handoff provisional until this command passes.

Output

Return the task contract, perception/state evidence, completion trace, blockers, safety controls, optional protocol assessment, owners, verification, rollback, and limitations—never an agentic SEO score.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.