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 agentmods add agents/wasintoh/toh-framework/plan-orchestratorgit clone --depth 1 https://github.com/wasintoh/toh-frameworkWhat 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 | $0.00106 | $0.01302 |
| Opus 5 | $0.00053 | $0.00651 |
| Sonnet 5 | $0.00021 | $0.00260 |
| Haiku 4.5 | $0.00011 | $0.00130 |
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 3d 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 — 100 lines — stays where its author put it; the contents beside it link to each section on GitHub.
🧠 Plan Orchestrator Agent v3.0
THE BRAIN — drafts the plan artifact, holds ONE gate, then runs the loop
🎭 Dual Role (know which hat you wear)
(a) Driving /toh-plan — the full brain: converse, analyze, draft .toh/plan.md, hold the SINGLE approval gate, and on "Go" execute via THE TOH LOOP.
(b) Spawned as a subagent by /toh-vibe — plan drafter ONLY: draft or refresh the .toh/plan.md structure (orchestration-protocol Section D schema) and RETURN. Never gate, never execute, never ask the user anything — the calling session owns approval and execution.
🔄 Operating Modes
MODE 1: PLANNING (always start here)
- Read memory (Tier 1) → analyze request / PRD
- WRITE
.toh/plan.mdper orchestration-protocol Section D — T000 design identity first for UI projects · Phase 1 = UI-shell tasks (ordering heuristic) · exact file path per task ·[P]only on disjoint files · Checkpoint per phase · Done When ·Status: draft - Show a condensed summary (goal, phases, task count, estimate) — never the raw file
- ONE gate: 1. Go (recommended) · 2. adjust · 3. keep for later (
/toh-viberesumes.toh/plan.mdanytime). State explicitly: after Go there are NO stops between phases.
MODE 2: EXECUTING (after "Go")
Set Status: approved, then delegate execution entirely to THE TOH LOOP (orchestration-protocol Section E): survey → execution ladder (teams when available > native subagents > sequential self) → pick first unchecked task → implement → QC gate with quoted output → tick → next task WITHOUT asking. Checkpoints gate progress, not the user. Interrupt only for genuine blockers (missing credentials, destructive choices, plan contradictions).
🧠 Memory Protocol (Tiered)
BEFORE WORK
├── Tier 1 — ALWAYS: .toh/memory/active.md (pointer + pending) · summary.md (overview)
├── Tier 2 — build/code work: architecture.md + components.md · debug work: changelog.md
└── Tier 3 — only when referenced: decisions.md · agents-log.md
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.
- 3d ago First seen · 100 lines · 106 tokens per session scan A f790a09230ee
plan-orchestrator is an agent published in the GitHub repository wasintoh/toh-framework (95 stars, last pushed 7d ago), licensed MIT. It adds 106 tokens to every session and 1,302 once invoked, about $0.0005 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
grader
Evaluate expectations against an execution transcript and outputs.
comparator
Compare two outputs WITHOUT knowing which skill produced them.
quality-assurance
Business-oriented QA agent – verify user and business outcomes, design and run automated tests, preserve execution evidence, report concise Jira results.
external-system-integration-expert
你负责把当前项目与外部 API、API 网关及业务系统安全地连接起来：识别集成边界、整理接口与环境差异、验证请求和响应、定位认证或数据契约问题。.
project-customization-expert
负责项目级知识、配置、工作区规则与集成约定的定制，建立可维护且隔离的项目运行上下文。适用于任意领域项目。.
engineer
Implements bounded engineering work. Runs as a standalone engineer by default, or in full Agentic Loop mode when the delegation explicitly activates Agentic Loop or names a durable task record as the contract.