I don’t get it.

“ Dependency Resolution: The harness resolves the DAG (Directed Acyclic Graph) automatically. No more relying on an LLM to "guess" the next logical step.Uses the Target: Dependency + Recipe model to ensure Agents follow a strict execution order without skipping steps.”

How does it do that? Does it just generate a Makefile? If so, why not just put the actual Makefile as a resource in the skill package and provide execution commands? That way the Makefile doesn’t need to be read at all.

If not, and you rely on an LLM interpreting the execution order, wouldn’t that statement just be false?

▲

hrimfaxi 2 hours ago | parent | next [-]

It seems like it relies on an LLM to guess the next logical step and codify it.

https://github.com/Teaonly/SKILL.make/blob/06872841537273376...

▲

teaonly 3 hours ago | parent | prev [-]

What I did here was to rewrite SKILL.md in Makefile style, using a DAG structure and omitting the text describing the process. So this should be considered a pseudo-Makefile; writing a SKILL using the Makefile approach is a very natural method.

▲

SwellJoe 3 hours ago | parent | next [-]

You're just repeating the readme, not answering the question.

▲

teaonly 3 hours ago | parent [-]

My next step is to design the recipe to be hot-loadable. The goal is to achieve self-evolving, optimizing the recipe independently without changing the DAG. This ability to perform local optimization is something Markdown lacks, but Makefiles can.

	▲	SwellJoe an hour ago \| parent \| next [-]
		I'm not sure using LLMs is good for your mental health. You should probably step away from the computer for a little while. LLMs are not always safe to use for everyone for reasons that I don't think are well understood, and I don't know exactly why it only causes trouble for some people, but the way you're talking is concerning. I'm being sincere here, not trying to be dismissive.
	▲	flexagoon 2 hours ago \| parent \| prev [-]
		How is this relevant to the question?

▲

stingraycharles 2 hours ago | parent | prev [-]

How is the DAG enforced, if not by executing “make” ? Then you’re just relying on the LLM to infer intent, which invalidates the claim, right?