wave-orchestration SKILL.md skillA
wave-orchestration from andrewbrown-usda/aigb-skills-usda is an agent skill: a SKILL.md folder an AI agent loads when a task calls for it.
mdr add andrewbrown-usda/aigb-skills-usda/wave-orchestration@v1.1Pins v1.1 in mdr.lock. Pin the fingerprint instead.
- Audit grade
- A, 17 of 17 checks passed
- Kind
- SKILL.md skill
- Repository
- andrewbrown-usda/aigb-skills-usda
- Path
skills/wave-orchestration/SKILL.md- Versions recorded
- 7
- Latest version
- v1.1, committed Oct 14, 2026
- First version recorded
- Sep 25, 2026
- Size
- 10,582 bytes
- Repository stars
- 7
- License
- a license file GitHub does not identify
Decomposes a large implementation plan into small-model-sized tasks executed by supervised worker agents in dependency-ordered parallel waves, with a plan-first approval gate inside every task, file-based state for context compaction, and a hard escalation ladder. Use when a master plan or multi-workstream feature is too large for one session and should be executed cheaply by many small agents under frontier-model supervision.
Collected from public GitHub. Every version is saved exactly as it was and never changes. Install it at an exact version with the mdr tool, and every file is checked before it reaches your agent. The main file is checked against the SHA-256 fingerprint recorded here, and the others against the fingerprints GitHub keeps for its source commit. The automatic checks below grade the latest version, and the same checks always give the same file the same grade.
What the file says
# Wave Orchestration (plan-gated, small-model workers) An architecture for executing a large, multi-workstream plan as a pipeline of small, cheap worker agents supervised by one frontier-model orchestrator. Three ideas do all the work: 1. **Waves over a task DAG** -- tasks grouped into coherent work clusters; tasks within a wave touch disjoint files and can run in parallel; waves are sequential and depend on prior waves. Every orchestration (planning-file set) can contain multiple waves. 2. **Plan-first inside every task** -- a worker may not write code until its micro-plan is approved. The approval gate is the *orchestrator*, not the human, so the loop stays autonomous. 3. **State on disk, not in context** -- a `STATE.md` file is the single resume point. Worker context never enters the orchestrator; the orchestrator's own context is disposable after every wave (compact or clear freely). ## Key terms - **Orchestration**: one planning-file set (EXECUTION_PLAN_N, PROMPT_TEMPLATE_N, STATE_N). Each orchestration has a unique number; wave numbers reset to 1 for each new orchestration. …
Read the whole file at its exact version.
How to install
mdr add andrewbrown-usda/aigb-skills-usda/wave-orchestration@sha256:74a0eda792e21c89The line at the top of this page pins the label v1.1, which is what most people want. This one pins the sha256 fingerprint of the contents instead, so it survives a label being moved. Either way the exact version is written to mdr.lock, and mdr install downloads and checks those contents again, so it installs them exactly or stops with an error.
[](https://markdownregistry.com/a/art_ddnt5emxp2z2iro6)
2 badge image fetches in 30 days. What the badge shows.
Versions
The registry has recorded 7 versions of this file, between Sep 25, 2026 and Oct 14, 2026, about one every 3 days.
| version | committed | commit | size | audit | |
|---|---|---|---|---|---|
| v1.1 latest | 2026-10-14 | 2620635 | 10,582 B | A | view · diff |
| v1.1 | 2026-10-14 | 677be34 | 10,676 B | A | view · diff |
| v1 | 2026-10-09 | 0d32dbb | 10,424 B | A | view · diff |
| v1 | 2026-10-02 | 511c0e3 | 10,424 B | A | view · diff |
| v1 | 2026-10-02 | 8c813e9 | 10,401 B | A | view · diff |
| v1 | 2026-09-30 | b8f4ba9 | 10,540 B | A | view · diff |
| v1 | 2026-09-25 | 160ae7c | 10,334 B | A | view |
Audit of the latest version
- pass: Frontmatter block present
- pass: Frontmatter declares a name
- pass: Frontmatter declares a description
- pass: Size between 200 bytes and 200 KB (10582 bytes)
- pass: No zero-width or bidi control characters
- pass: No instruction hidden inside an HTML comment
- pass: No link to an exfiltration or paste host
- pass: No credential-shaped string
- pass: No instruction to send local credentials anywhere
- pass: No text hidden with inline styles
- pass: No prompt-injection phrasing
- pass: No curl or wget piped into a shell
- pass: No recursive delete of root, home or parent
- pass: No instruction to read or print local credentials
- pass: No base64 blob over 200 characters
- pass: No link to a raw IP address
- pass: No script tag
Source
andrewbrown-usda/aigb-skills-usda · 7 stars · license a license file GitHub does not identify · pushed 2026-10-21 · branch main
API
GET https://markdownregistry.com/api/v1/artifacts/art_ddnt5emxp2z2iro6 GET https://markdownregistry.com/api/v1/resolve?ref=andrewbrown-usda/aigb-skills-usda/wave-orchestration GET https://markdownregistry.com/api/v1/blob/74a0eda792e21c8927e3149e67d1f6c3c86f8ce1c3ff0df55a581af002d213b7
Your AI assistant does the legwork. You hear about the deals worth your word. Hand yours the standing instructions at modelranch.com and it joins the network that reads files like this one.
Guides: the SKILL.md format and its required fields · what to check before you install a skill · how to pin a skill to an exact version · where to find agent skills · keeping a team on the same skills · which agent file does what.