agent-fanout
Skill0 starsChanged 25 days agoMITView on GitHub
Task-scoped parallel worker waves for explicitly requested fanout or clearly separable multi-part work. Do not use for simple answers, single small edits, private tools, secrets, git mutation, or destructive actions.
1 other project uses a version of this.
Adopted by
Most recently taken up by
- audriusstat/execution-disciplinetheir own version, first seen Jul 28, 2026
Other versions1 project uses a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 1
- audriusstat/execution-discipline★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repositoryOne other file it gives its agents.
- adapters/Claude.mdCLAUDE.md
The file
We split this skill into the rules it gives an agent
10 rules. About 0 are common, 6 are distinctive, and 4 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈6
Distinctive · estimate
This author's own choices. Worth a look.
4
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 9
If it is not useful, continue normally and silently.
- Very specific: line 10
Use task-scoped AgentFanout <task>, AgentFanout Codex|Grok|Local <task>, Help, Status, or Cancel.
- Distinctive: line 11
Treat GPT-6 Astra as an explicit cost escalation: never select it automatically, never launch more than one Astra worker in a wave, and …
- Very specific: line 12
Prefer native harness agents.
- Distinctive: line 14
Use a verified host lease when available.
- Distinctive: line 15
Keep private tools, secrets, git mutation, destructive actions, and final synthesis in the parent.
- Distinctive: line 18
After every result batch, rebuild the residual ledger and launch or follow up the next wave while delegable work remains.
- Distinctive: line 19
Follow up with the same worker only for unchanged scope.
- Distinctive: line 20
Use validators only for risk, contradiction, or an explicit request.
- Very specific: line 22
Do not claim model control, enforcement, portability, or savings without host evidence.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
AgentFanout
Activate only for an explicit fanout request or clearly separable parallel work.
If it is not useful, continue normally and silently; do not mention AgentFanout.
Use task-scoped AgentFanout <task>, AgentFanout Codex|Grok|Local <task>, Help, Status, or Cancel.
Treat GPT-6 Astra as an explicit cost escalation: never select it automatically, never launch more than one Astra worker in a wave, and require an exact-model launch acknowledgement plus an authorization reason.
Prefer native harness agents; use the optional hardened runtime only when explicitly requested.
For an enabled V2 candidate, load workflows/V2.md; dispatch only its acknowledged assignment plan, with Codex-only automatic routing.
Use a verified host lease when available; otherwise label orchestration advisory and never claim enforcement.
Keep private tools, secrets, git mutation, destructive actions, and final synthesis in the parent.
Split work into the smallest bounded, non-overlapping task packets and launch independent work concurrently.
Ingest only compact findings with evidence, uncertainty, and residual work; never paste raw transcripts or route JSON.
After every result batch, rebuild the residual ledger and launch or follow up the next wave while delegable work remains.
Follow up with the same worker only for unchanged scope; use a fresh worker for changed scope or independent validation.
Use validators only for risk, contradiction, or an explicit request; do not add default low-risk validation.
Handshake covers spawn, wait, follow-up, cancel, callback, receiver ID, model/effort, and enforcement. Require host IDs.
Do not claim model control, enforcement, portability, or savings without host evidence; benchmarks govern savings claims.
Load workflows/Activate.md for selection and workflows/Wave.md for active multi-wave execution.
Load references/ParentBoundary.md and references/WhenNotToFanout.md only when needed.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.