Getting it into your agent
This one installs as part of its plugin. Adding the marketplace and installing the plugin brings it with everything else the plugin ships.
/plugin marketplace add shalintripathi/organic-os/plugin install organic-osWrote 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/skills/shalintripathi/organic-os/hoo-orchestrator)<a href="https://agentmods.dev/skills/shalintripathi/organic-os/hoo-orchestrator"><img src="https://agentmods.dev/badge/skills/shalintripathi/organic-os/hoo-orchestrator/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/skills/shalintripathi/organic-os/hoo-orchestrator"><img src="https://agentmods.dev/badge/skills/shalintripathi/organic-os/hoo-orchestrator.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 | $0.00056 | $0.00998 |
| Opus 5 | $0.00028 | $0.00499 |
| Sonnet 5 | $0.00011 | $0.00200 |
| Haiku 4.5 | $0.00006 | $0.00100 |
Grade A, and why
hoo-orchestrator scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 11d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
What it actually says
Head of Organic - orchestrator
- Locate the brain repo: try
core.registry.get_active()first (the active site'sbrainpath). If the registry is empty or unavailable (cloud runtimes may lack local registry access), fall back to the routine's own configured site-profile path, or ask the user (a repo has site-profile.yaml at root). Read site-profile.yaml fully. 1a. Runcore.contracts.check_schema(brain_path)before fanning out to any specialist. If not compatible, relay the action string and stop. - Decide which specialists the request needs (default full sweep: all eight). Launch them as parallel agents, each given the profile path + target URLs.
- Synthesize results. Deduplicate findings. Rank by impact x confidence.
3b. Decision check, BEFORE any
create_itemcall. For every brief topic and every on-page fix, search the brain's decision memory:PYTHONPATH="$CLAUDE_PLUGIN_ROOT/lib" python3 -c "..."snippet importingcore.decisions-search(<brain>, [<page path>, <topic or keyword>]). It returns prior decisions newest first withdate,choice,scope,item, and therationale. A hit whosechoiceisrejectedmeans a human already refused that work. Exactly two paths are allowed, never a third:- SKIP it, and say so in step 6's summary: name the topic, the decision date, and the recorded reason.
- Or CREATE it with a line in the item body reading "previously rejected
on because ; proposing again because ".
What changed must be a concrete new fact from this run - a new signal,
a ranking move, a competitor change - never "worth another look".
Silently re-proposing rejected work is forbidden: an unread rejection is
how a loop wastes a human's attention twice. A hit with any other
choiceis context to quote in the body, not a blocker.
- File outputs through lib/core ONLY:
- observations ->
append_signal(one call per signal line) - content ideas ->
create_item(kind="content-brief", ...); ideas born from comparison-intent queries (vs, alternative, best X for) setbrief_type="comparison"- absence means explainer - on-page fixes ->
create_item(kind="onpage-fix", ...) - run artifacts ->
runs/YYYYMMDD-orchestrator/(numbered raw files + REPORT.md) Items are created ONLY viacreate_item(bornproposed, empty approvals) or the contract CLI - never by writing abriefs/orproposals/file directly, and never with any other status at creation. A brief that ce-produce will later draft still startsproposedand reachesdraftedonly through approved lineage.
- observations ->
- Rebuild the queue (
rebuild_queue) and notify per the profile's approval channel (see skills/onsite-apply for the adapter pattern) with one call:PYTHONPATH="$CLAUDE_PLUGIN_ROOT/lib" python3 -c "..."importingcore.approvaland callingnotify_pending(<brain>, send), wheresenddelivers over that channel. It sends every un-notified proposed item and marks each as it goes; an item is marked notified ONLY after its send returns without raising, so a failed send is retried on the next run rather than lost, and re-runs never re-notify. Do not callis_notifiedormark_notifiedby hand. Do NOT apply anything: creating items is free, mutating the site is gated elsewhere. - Tell the user: top 5 actions, what is queued for approval, what was skipped for missing credentials.
Every signal line must be falsifiable: observation + "we are wrong if" + a leading indicator. Reject vague signals.
At each stage boundary, append a one-line progress marker with a UTC timestamp to the run report file before starting the stage - headless runs are watched by tailing that file, not a terminal.
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.
- 11d ago First seen · 66 lines · 56 tokens per session scan A 1487a2a2a598
hoo-orchestrator is a skill published in the GitHub repository shalintripathi/organic-os (5 stars, last pushed 3d ago), licensed MIT. It adds 56 tokens to every session and 998 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 2026-08-31.
Other skills, from other repositories
seo-vertical-detect
Classify a website as ecommerce, local-business, blog-publisher, saas, docs, or generic (multiple may apply), and report detected locales, from the persisted PageSnapshot plus the platform detector's vertical hints. Used by seo-orchestrator to decide which conditional modules to run and how to reweight the scores.
seo-fix-apply
Writer protocol preloaded into the seo-fixer-writer agent — how confirmed SEO/AI-search fixes are applied (git pre-flight, backup, Edit/Write for local diffs, ticketed adapter CLIs for remote targets, re-verify, publish only on a second ticket, rollback) and the findings-array output contract. Used only by the fix…
audit
Audit a website or web codebase for SEO and AI-search (GEO/AEO) — produces two independent 0-100 scores (Search SEO + AI Visibility) plus a prioritized, evidence-backed report persisted on disk. Read-only; never writes to the project. Use when the user asks to audit, analyze, check, or score a site's SEO, structured…
ai-visibility
Fast-path AI visibility — get a brand's 0–100 Akii Visibility Score (computed by an open-source LLM judge against the brand's public footprint) with four-dimension breakdown AND a per-engine proxy map for ChatGPT, Claude, Gemini, Perplexity, Copilot, and Google AI Overviews based on FirstPageSage signal weights.…
optimize-page
Comprehensive single-page optimization across all three layers — traditional SEO (title / meta / H1 / internal links), AEO (chunk quality, direct-answer leads, FAQ extraction), and GEO rewrites using the tactics published by the Princeton/IIT Delhi GEO study (citation integration, expert quotes, statistics enrichment…
seo-audit
Single audit entry point — surface-level scorecard or deep infrastructure dive depending on the requested mode. Default full mode produces a multi-layer scorecard across all 9 areas (crawlability, indexation, meta tags, headings, images, Core Web Vitals, JS rendering, mobile + security, structured data, internal…