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.
npx skills add wasintoh/toh-framework --skill plan-orchestratorgit clone --depth 1 https://github.com/wasintoh/toh-frameworkWrote 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/wasintoh/toh-framework/plan-orchestrator)<a href="https://agentmods.dev/skills/wasintoh/toh-framework/plan-orchestrator"><img src="https://agentmods.dev/badge/skills/wasintoh/toh-framework/plan-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/wasintoh/toh-framework/plan-orchestrator"><img src="https://agentmods.dev/badge/skills/wasintoh/toh-framework/plan-orchestrator.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00094 | $0.01190 |
| Opus 5.5 | $0.00038 | $0.00476 |
| Sonnet 5.5 | $0.00019 | $0.00238 |
| Haiku 4.5 | $0.00009 | $0.00119 |
Grade A, and why
plan-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 1mo 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.
How it starts
The opening of the file, as written. The whole thing — 82 lines — stays where its author put it; the contents beside it link to each section on GitHub.
🧠 Plan Orchestrator Skill v3.0
Planning + plan artifact + single-gate handoff to autonomous execution For the Plan Orchestrator Agent · drives
/toh-plan· drafts plans for/toh-vibe
🎯 Core Philosophy
THE BRAIN of Toh Framework. Two jobs, cleanly split:
- PLAN — converse, analyze, and produce
.toh/plan.md(a file, never chat state) - HAND OFF — one "Go" gate, then execution belongs to THE TOH LOOP (orchestration-protocol)
🔄 MODE 1: PLANNING (default)
When receiving /toh-plan:
- Read memory (Tier 1:
active.md+summary.md; deeper tiers per task type) - Analyze the request / PRD — business type, users, features that actually matter
- WRITE
.toh/plan.mdper the orchestration-protocol Section D schema — that section is the single source of truth for the format; never invent your own. Honor in particular:- Task grammar
- [ ] T001 [P] agent-name — description in exact/path.tsx T000 design-reviewer — generate root DESIGN.mdfirst for any UI project- Checkpoint per phase · Done When for the whole plan ·
Status: draft
- Task grammar
- Show the user a condensed summary (goal, phases, task count, estimate) — not the raw file, no giant tables
- Hold the ONE approval gate: 1. Go (recommended) · 2. adjust · 3. keep for later — and state explicitly that after Go there are no per-phase stops
While planning: adjust ("add X", "cut X") → edit plan.md, re-summarize · questions → answer freely · "Go" → MODE 2.
🔄 MODE 2: EXECUTING
One line: execution = THE TOH LOOP (orchestration-protocol Section E); a single "Go" replaces all per-phase confirmation. Set Status: approved and run it — checkpoints gate progress, not the user.
🧩 Phase Design Heuristics
- UI-first ordering (a heuristic encoded in the plan, not a per-phase rule): Phase 1 tasks are UI-shell tasks so the user sees screens early; logic/backend phases follow. Don't force UI tasks into every phase.
- Small tasks: each task fits one context window and names its exact file path.
[P]parallel-safe only when files are disjoint.- Realistic estimates: simple page 1-2 min · page with forms 3-5 · complex logic 5-10 · schema/RLS 2-3 · design polish 3-5. Over-estimate rather than under-deliver.
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.
- 1mo ago First seen · 82 lines · 94 tokens per session scan A 8cda4fa2497e
plan-orchestrator is a skill published in the GitHub repository wasintoh/toh-framework (98 stars, last pushed 1mo ago), licensed MIT. It adds 94 tokens to every session and 1,190 once invoked, about $0.0004 per session on Opus 5.5. 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 skills, from other repositories
goga-task-by-specifying
Interactive task formulation from a raw request.
goga-specify
User task formulation specialist.
task-record-contract
Use when creating, updating, or reading the task record for a single implementation task. Covers required task sections, no-placeholder rules, files-first task-file frontmatter, work-unit summary template, implementation artifact linkage, and safe GitHub projection posting practices. Backend projections are documented…
setup-agenticloop
Use when a maintainer or human is setting up Agentic Loop in a target project for the first time, when source document names are non-standard or ambiguous, when .agenticloop/project.md is still unconfirmed, or when a human wants to update its development-stage profile. Discovers bounded candidate project docs once…
prepare-ticket
Use when asked whether a ticket or issue is ready for development, or to prepare, validate, refine, challenge, or sanity-check a Jira work item, GitHub Issue, or GitLab Issue before building it. Checks material claims against evidence, follows questions until outcome decisions are settled, rewrites the title and…
implement-ticket
Use when asked to implement, build, or pick up a ticket end to end — a Jira work item, GitHub Issue, or GitLab Issue — including "implement ABC-123", "implement gh issue #137", "work on issue 137", or "pick up this ticket". Critiques the request against the code, builds it in a throwaway git worktree, verifies it with…