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 skills/davidomil/cloudx/change-orchestratornpx skills add davidomil/cloudx --skill change-orchestratorgit clone --depth 1 https://github.com/davidomil/cloudxWrote 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/davidomil/cloudx/change-orchestrator)<a href="https://agentmods.dev/skills/davidomil/cloudx/change-orchestrator"><img src="https://agentmods.dev/badge/skills/davidomil/cloudx/change-orchestrator.svg" alt="Measured on agentmods" 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 | $0.00031 | $0.00488 |
| Opus 5 | $0.00015 | $0.00244 |
| Sonnet 5 | $0.00006 | $0.00098 |
| Haiku 4.5 | $0.00003 | $0.00049 |
Grade A, and why
change-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 4d 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 — 53 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Change Orchestrator
Responsibility
Route one change through the repository state machine. Do not plan, implement, verify, review, or mutate GitHub in the orchestrator context.
Inputs
- Original task and current worktree/base SHA.
- Root and scoped
AGENTS.mdfiles. .agents/pr-review-policy.tomlclassification.- Current typed artifacts and finding IDs, if resuming.
Workflow
- Classify exact proposed paths and type through
scripts/ai-change/policy.mjs. - Start a fresh
$plan-changecontext and validate its output against.agents/schemas/plan.schema.json. - Start a different fresh
$review-plancontext. A blocked review returns to a new planner; a review ceiling transitions toblocked. - Dispatch
$implement-changeonly after a clean plan review. - Dispatch
$verify-change. Verification cannot edit or fix the worktree. - Start fresh contexts for every reviewer selected by policy. Give each only the task, trusted instructions, accepted plan, implementation artifact, current diff, and verification artifact.
- Dispatch
$review-changeto reconcile the exact-head findings. Findings return to a fresh implementer, followed by full verification and review. - After a clean local gate, dispatch
$review-prfor the current PR head. - Dispatch
$ship-changeonly when the user explicitly requests a GitHub mutation or merge action.
The private AI manager owns durable lifecycle transitions. This public skill produces schema-valid artifacts and exact-head evidence for that manager; it never infers a later state from prose. An iteration cap, missing reviewer, stale SHA, schema error, or failed command blocks the run and never permits a bypass.
Output
Return the current state, artifact paths/digests, finding IDs, exact commands run, and the next valid action. Do not produce a substitute artifact in prose.
Prohibitions
- No author self-review or reused author conversation.
- No code edits, commits, pushes, labels, PR reviews, approvals, or merges.
- No partial PR after failed review or verification.
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.
- 4d ago First seen · 53 lines · 31 tokens per session scan A 0bf3edcdf262
change-orchestrator is a skill published in the GitHub repository davidomil/cloudx (1 stars, last pushed 11d ago), licensed MIT. It adds 31 tokens to every session and 488 once invoked, about $0.0002 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
simply-pm
Use when you are spawned as the PM of a simply harness project and your brief references a .simply/LEDGER.md work ledger or this skill. Grants fleet authority, routing doctrine, safe worker lifecycle, and the signal protocol back to the auditor.
authoring-prd
Use when turning a product idea or change into a loop-ready continuous backlog with small stories and executable behavioral evidence.
project-contract-harness
Durable contract harness for long-running large-project delivery by coding agents. Use when an agent must drive a new project, takeover, module, feature, or complex repo through project analysis, persistent local state, project-specific runners, secret-access rules without storing secrets, centralized artifact…
product-manager
Use when defining product requirements, prioritizing features, planning a roadmap, validating user problems, or making build/buy/don't-build decisions.
claude-orchestrator
Coordinate bounded Claude implementation workers with durable authenticated A2A/LAN jobs, explicit evidence, recovery, and read-only verification.
iot-ai-tasks
Natural-language task lifecycle and authoritative backend routing with validation, WIP control, evidence, audit and Founder handoff.