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 joslat/maf-doctor --skill maf-fan-out-validatorgit clone --depth 1 https://github.com/joslat/maf-doctorWrote 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/joslat/maf-doctor/maf-fan-out-validator)<a href="https://agentmods.dev/skills/joslat/maf-doctor/maf-fan-out-validator"><img src="https://agentmods.dev/badge/skills/joslat/maf-doctor/maf-fan-out-validator/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/joslat/maf-doctor/maf-fan-out-validator"><img src="https://agentmods.dev/badge/skills/joslat/maf-doctor/maf-fan-out-validator.svg" alt="Reviewed on agentmods" width="80" 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 5.1 | $0.00055 | $0.00828 |
| Opus 5.5 | $0.00022 | $0.00331 |
| Sonnet 5.5 | $0.00011 | $0.00166 |
| Haiku 4.5 | $0.00006 | $0.00083 |
Grade A, and why
maf-fan-out-validator 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 2mo 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 — 56 lines — stays where its author put it; the contents beside it link to each section on GitHub.
fan-out-validator — silent fan-in starvation detection
⚡ Prefer the MCP tools. The
maf-autopilotMCP server exposes:
MafValidateFanOut(repoPath)— Roslyn syntax walk; flags any[MessageHandler]returningvoid/ non-genericTask/ValueTask(the silent-starvation patterns).MafSimulateWorkflow(repoPath)— builds the full executor topology graph (fan-out + fan-in edges resolved through variable bindings), emits a Mermaid diagram, and gives an end-to-end completion forecast.Use both:
MafValidateFanOutfor per-handler verdicts,MafSimulateWorkflowfor "would this workflow complete?"This skill formerly had a sibling
fan-in-static-analyzerthat walked the same patterns by hand; both have been merged here now that the MCP tools cover the procedural work.
The bug class — "silent fan-in starvation"
A fan-out executor's [MessageHandler] method MUST return one of:
Task<TMessage>ValueTask<TMessage>IAsyncEnumerable<TMessage>(streaming)
A handler that returns void, Task, or ValueTask (non-generic) produces no output message. The fan-in barrier on the downstream edge then starves silently — the workflow exits cleanly but incompletely.
This is NOT a build error. dotnet build is clean. Tests that don't exercise the fan-in path are clean. The bug only surfaces in production traces (or with one of the tools above).
Return-type verdict table
| Return type | Verdict | Why |
|---|---|---|
Task<TMessage>, ValueTask<TMessage> |
✅ OK | Produces a downstream message |
IAsyncEnumerable<TMessage> |
✅ OK (streaming pattern) | Produces a stream of messages |
void |
❌ SILENT_STARVATION_RISK | No completion signal AND no message |
Task, ValueTask (non-generic) |
❌ SILENT_STARVATION_RISK | Completes but emits nothing |
int, string, raw concrete types |
⚠ LIKELY_INVALID | Cannot satisfy MAF's fan-out contract |
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.
- 2mo ago First seen · 56 lines · 55 tokens per session scan A 1d1ac84e0f4a
maf-fan-out-validator is a skill published in the GitHub repository joslat/maf-doctor (14 stars, last pushed 3d ago), licensed MIT. It adds 55 tokens to every session and 828 once invoked, about $0.0002 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
eve
Build durable backend AI agents with the eve framework. Use when creating, editing, or debugging an eve project — agent instructions, skills, tools, connections, channels, sandboxes, subagents, schedules, or evals.
investigate
Investigate a specific issue report with strict, evidence-driven diagnosis.
bfl-api
BFL FLUX API integration guide covering endpoints, async polling patterns, rate limiting, error handling, webhooks, and regional endpoints with Python and TypeScript code examples.
windbg-diagnostic-method
Use with every WinDbg plugin investigation to apply evidence-first reasoning, confidence calibration, contrarian review, structured reporting, and deterministic validation. Not a bug-family-specific triage skill.
windbg-kernel-bugcheck-triage
Use when a kernel dump reports a Windows bugcheck; decode parameters and recover exception or trap context before investigating your driver. Not for user-mode process crashes or blaming a module from its name alone.
windbg-kernel-irp-lifecycle-triage
Use when kernel evidence shows stalled I/O, a power IRP, or completion/cancellation misuse; inspect request state and driver ownership. Not for interpreting an empty IRP search in a limited dump as proof of healthy I/O.