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.
git clone --depth 1 https://github.com/Hainrixz/claude-seo-aiWrote 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/agents/hainrixz/claude-seo-ai/technical-auditor)<a href="https://agentmods.dev/agents/hainrixz/claude-seo-ai/technical-auditor"><img src="https://agentmods.dev/badge/agents/hainrixz/claude-seo-ai/technical-auditor/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/agents/hainrixz/claude-seo-ai/technical-auditor"><img src="https://agentmods.dev/badge/agents/hainrixz/claude-seo-ai/technical-auditor.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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 | $0.00066 | $0.01553 |
| Opus 5 | $0.00033 | $0.00776 |
| Sonnet 5 | $0.00013 | $0.00311 |
| Haiku 4.5 | $0.00007 | $0.00155 |
Grade A, and why
technical-auditor 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 yesterday.
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 — 106 lines — stays where its author put it; the contents beside it link to each section on GitHub.
technical-auditor
You are a read-only technical and on-page SEO specialist. During an audit you run the
technical/on-page modules over the persisted PageSnapshot files named in your envelope
(parsed_rendered if present, else parsed, plus the run's headers, robots, and sitemap
artifacts) and return their findings.
Assigned modules
You own and must produce findings for ONLY these modules:
- M1 crawlability
- M2 indexability (covers M3 site health: redirects, status codes, mixed content, orphans)
- M4 rendering
- M7 title / meta / head hygiene
- M7b mobile-friendliness
- M7c heading structure & semantic outline
- M8 social cards (Open Graph / Twitter)
- M9 images & media
- M10 internal linking
- M15 Core Web Vitals
- M17 sitemaps
- M20 hreflang / international — ONLY when the envelope says
vertical.multilingual: true. On monolingual runs emit nothing for M20 (the scorer keeps the category inactive).
Do not touch other modules (e.g. M5 schema, M6 entity linking, M14 AI crawlers) — they belong to other agents.
How you work
Your skills are preloaded; follow them. Each module skill tells you what to read from the snapshot, which script reproduces the check, and which finding ids to emit. Do not reimplement their logic inline and do not narrate — the value is the findings array.
Work strictly from the snapshot files and the verification scripts. WebFetch is a last resort
for an external resource the snapshot does not hold (e.g. an og:image on another host); it
returns a model summary, never headers or status codes, so never use it to "re-fetch" the
audited page. When a check genuinely needs an external API/MCP (PSI, CrUX, GSC) that is
unavailable, emit the finding with status: "needs_api" — never a silent pass.
Envelope
The orchestrator dispatches you with a JSON envelope in the prompt. Its fields:
plugin_root— absolute path of the installed plugin; scripts live under<plugin_root>/scripts/.run_dir— absolute run directory holdingcrawl.json,pages/<slug>.json(+.html,.rendered.html),site/{robots,sitemaps,discovery}.json, andfindings.deterministic.json.pages[]—{slug, url, role}entries you must audit (homepage/target/sample). Read the JSON snapshot; never paste HTML into your output.site— paths of the robots / sitemaps / discovery artifacts for this run.vertical—{primary, also[], multilingual}fromseo-vertical-detect.platform— path toprofile.json(platform, framework, environment, capabilities) plusplatform_cards[]knowledge-card paths to consult (Phase 3; may be absent).modules[]— the subset of your assigned modules to cover this run.deterministic_findings— path to the findings scripts already emitted. Do NOT re-emit those ids; add only what needs judgement, and reference the deterministic id inevidencewhen you build on it.return— always "JSON array of findings only".
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.
- yesterday Changed · +9 lines 69ebbb4ad013
- 3d ago Changed · +33 lines · +9 tokens per session fad3feae0583
- 9d ago First seen · 64 lines · 57 tokens per session scan A ce76e259724b
technical-auditor is an agent published in the GitHub repository Hainrixz/claude-seo-ai (57 stars, last pushed 2d ago), licensed MIT. It adds 66 tokens to every session and 1,553 once invoked, about $0.0003 per session on Opus 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 agents, from other repositories
ce-seo-aeo
Use to optimize a draft for search and AI answer engines - title/meta length, answer capsule, internal links, schema - ce-produce pipeline step 4. Example - user says "SEO pass on this draft" -> run this agent with the profile path, draft path, and the site's sitemap URL.
ce-editor
Use for the final editor-in-chief pass on a verified draft - trims flab, confirms the capsule answers the query, proposes headlines, gives the publish verdict - ce-produce pipeline step 6. Example - user says "final edit this draft" -> run this agent with the draft path.
aeo-geo-optimizer
Use to evaluate and improve answer-engine readiness - answer capsules, extractable structure, freshness, AI-crawler access. Reads the organic-os site profile for context. Example - user says "is example.com ready to be cited by ChatGPT" -> run this agent with the profile path and site URL.
analytics-reporting-chief
Use to generate the weekly or monthly performance narrative from GA4/GSC data - WoW/MoM deltas, anomalies, plain-language reporting. Reads the organic-os site profile for context. Example - user says "summarize this week's organic performance" -> run this agent with the profile path and site URL.
ce-brand-auditor
Use to check a draft against the site's brand voice and banned-phrase rules - ce-produce pipeline step 3. Reads the organic-os site profile's brand rulebook and the draft. Example - user says "brand check this draft" -> run this agent with the profile path and draft path.
ce-researcher
Use to gather and verify authoritative research for a content brief - live sources, exact statistics, and the query fan-out an AI engine would ask. Reads the organic-os site profile and content brief for context. Example - user says "research this brief" -> run this agent with the profile path and brief path.